Poslední aktualizace: 2026.09.02 | Externí odkaz

Privacy Policy and Terms and Conditions for PyGPT Software

---

Interpretation and Definitions

Interpretation

Words with the first letter capitalized have specific meanings defined under these conditions. These definitions apply regardless of whether the terms appear in singular or plural form.

Definitions

For the purposes of this Privacy Policy and Terms and Conditions:

• "Author" refers to the creator(s) of PyGPT, also referred to as "We," "Us," or "Our" in this document.
• "App" or "Application" means the PyGPT software application.
• "You" or "User" means any individual or entity using the App or website.
• "Provider" means a third-party AI, API, cloud, model, search, or other service used or configured through the App.

---

Disclaimer

General Information

Information and content provided through the App are for general purposes only. The Author assumes no responsibility for errors, omissions, inaccuracies, or outdated information contained in responses generated by AI models or other external services.

AI-Generated Content

Responses generated by artificial intelligence models may be inaccurate, incomplete, misleading, or inappropriate for a particular purpose. You are responsible for verifying important information before relying on it, especially in professional, financial, legal, medical, security-related, or other sensitive contexts.

Warranty and Liability

The App is provided "as is," without warranty of any kind, express or implied. To the maximum extent permitted by applicable law, the Author shall not be liable for any direct, indirect, incidental, special, consequential, or other damages arising from the use of, or inability to use, the App.

Age Restrictions

The App is intended for users who are at least 13 years old. Individual Providers may impose different or additional age requirements. You are responsible for complying with the terms and age requirements applicable to the services and Providers you use.

External Links

The App and website may contain links to third-party websites and services that are not operated or controlled by the Author. We are not responsible for the content, availability, security, privacy practices, or policies of third-party websites or services.

Errors and Omissions

Information provided through the App should not be relied upon as the sole basis for making important decisions without consulting appropriate primary, authoritative, or otherwise reliable sources.

Use at Your Own Risk

The App and information made available through it are provided without guarantees of completeness, accuracy, reliability, availability, or usefulness. You use the App, its tools, integrations, generated content, and external services at your own risk.

---

API, Token and Usage Costs

PyGPT can connect to multiple AI and service Providers, including but not limited to OpenAI, Anthropic, Google, xAI, Perplexity, Hugging Face, local model providers, and other services configured by the User.

API usage, token consumption, quotas, rate limits, subscription requirements, and billing rules are determined independently by each Provider. PyGPT does not control the pricing, billing, token calculation, usage limits, or account policies of third-party Providers.

Certain App features, plugins, agents, tools, attachments, multimodal requests, retries, background operations, or multi-step tasks may result in additional API requests and therefore additional token usage or other charges.

You are responsible for monitoring your API usage and any associated costs. Always check the usage dashboard, billing information, pricing documentation, and terms provided by the Provider whose service you are using.

---

Privacy of Your Data

Local Data and Data Handling

PyGPT is a desktop application and stores its application data and configuration locally unless a particular feature, integration, or Provider requires data to be transmitted to an external service.

The Author does not receive your conversation content, prompts, generated responses, files, or API keys merely because you use the App. Limited technical information may be transmitted when required for features such as version checks, updates, or other explicitly enabled online functionality.

Third-Party AI Providers and External Services

PyGPT acts as a client for third-party AI models, APIs, and other external services. When you select or configure an external Provider, data necessary to perform your request may be transmitted to that Provider.

Depending on the feature being used, transmitted data may include prompts, conversation context, instructions, attachments, images, audio, files, tool inputs and outputs, or other content required to process your request.

The processing, storage, retention, logging, use, and possible use for service improvement or model training of data sent to a third-party Provider are governed by that Provider's own privacy policy, terms of service, API terms, data retention rules, and account settings.

Different Providers may apply different rules, and those rules may also differ between consumer services, commercial services, enterprise products, and APIs. Before using a Provider through PyGPT, you should review the current privacy policy, terms of service, data usage policy, and retention policy published on that Provider's official website.

This applies to services including, but not limited to: OpenAI, Anthropic, Google, xAI, Perplexity, Hugging Face, and any other third-party Provider, API, model endpoint, plugin, tool, or external service that you configure or use with PyGPT.

API Keys and Credentials

API keys and other credentials entered into PyGPT are used to authenticate requests to the corresponding Provider or service. They are not intentionally transmitted to the Author.

Your credentials may necessarily be transmitted to the relevant Provider as part of the authentication process when making API requests. You are responsible for protecting your API keys, configuring appropriate usage limits where available, and complying with the Provider's security and account requirements.

Plugins, Tools and Integrations

Some plugins, tools, integrations, web services, or optional App features may communicate with third-party services independently of the selected AI Provider. Data required to perform such operations may therefore be sent to those services.

You should review the privacy policies and terms of any external services that you enable or configure in the App.

Local Models and Services

PyGPT may also be configured to work with locally hosted models or services. When using a fully local configuration, data processing may remain on your local device, depending on the model, software, plugins, tools, and integrations you choose to use.

---

PyGPT.net Website Privacy Policy

Privacy Commitment

At pygpt.net, we respect the privacy of our visitors. This section describes the information that may be collected when you use the website and how that information is used.

Contact Form and Enquiries

When you contact us using the contact form, we process the information you provide, such as your name, email address, subject, and message, solely for the purpose of receiving, processing, and responding to your enquiry.

Technical information associated with the request, such as your IP address, may also be processed for security, spam prevention, abuse prevention, and technical troubleshooting purposes.

Information submitted through the contact form is not sold or used for unrelated marketing purposes. It is retained only for as long as reasonably necessary to process your enquiry and for legitimate administrative, security, or legal purposes where applicable.

Queries and Contact

For questions concerning this Privacy Policy or the processing of your data, please contact us at [email protected].

Log Files and Tracking

Like many websites, pygpt.net uses standard server log files. These logs may contain technical information such as IP address, browser type, operating system, referring page, requested resources, timestamps, and similar diagnostic information.

Such information may be used for website operation, security, troubleshooting, statistical analysis, and prevention of abuse.

Cookies and Third-Party Services

The website may use cookies or similar technologies where required for website functionality, user sessions, statistics, security, or services provided by third parties.

Some website functionality may rely on services provided by Google, including services used for statistics, security, or protection against automated abuse. Google's handling of information is governed by its own privacy policies and terms.

Google's Privacy Policy: https://policies.google.com/privacy

Children's Online Privacy

Protecting children's privacy online is important. We encourage parents and guardians to monitor, participate in, and guide their children's online activity.

Scope of Website Privacy Policy

This website privacy section applies to information processed in connection with the use of pygpt.net. Data processed by third-party Providers through the PyGPT desktop application is governed separately by the applicable Provider's policies, as described above.

---

License

The Application is provided under the MIT License.

MIT License

Copyright (c) 2026 Marcin Szczygliński

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

---

Your Acknowledgement

By using the App or website, you acknowledge that you have read and understood this Privacy Policy and the applicable Terms and Conditions. Your use of third-party Providers and services is additionally subject to their respective terms, privacy policies, and other applicable rules.

Seznam změn

2.9.1 (2026-10-06)

- Improved attachment uploads and handling; added remote connectors to the Add attachment menu, shared context between chats within a project, and @mentions for referencing other chats in the same project.
- Merged the Files I/O, Python Interpreter, and System (OS) plugins into a single integrated plugin: Filesystem, Python and OS.
- Added @mention search by chat title alongside exact ID lookup.
- Added configurable toolbox placement: left, middle, or right.
- Added Tab completion support in the terminal.
- Optimized directory discovery for file searches in @mentions.
- Fixed various UI issues.

2.9.0 (2026-10-04)

- Modernized the layout: moved the toolbox to the left, hidden it by default, and added a dynamic right column.
- Added a toolbar on the left.
- Added a new Terminal tool.
- Added pending input submission during generation, with interruption and continuation of active turns.
- Refined agent instructions, execution workflows, and progress reporting.
- Improved UI and UX.
- Simplified attachment uploads.
- Refactored the tools core.
- Improved tab management and added support for opening new browser tabs.
- Optimized layout updates.
- Added many other improvements and fixes.

2.8.38 (2026-10-03)

- Improved attachment handling and attachments display.
- Optimized and improved Computer Use, fixed the Computer Use flow in agents, and added a full-screen indicator while Computer Use is active.
- Added support for Markdown (.md) files in the Files preview.
- Added a new File Preview add-on type, allowing support for additional file formats to be extended through add-ons.
- Added GPT-6.1 Sol model.
- Added a Group models by provider option.
- Added a Show tools in agents option.
- Other UI/UX improvements and fixes.

2.8.37 (2026-10-01)

- Added SHA-256 integrity verification for public trusted Add-ons, including manifest/registry hash pinning and validation during installation.
- Added helper scripts in bin/ for generating SHA-256 signatures for community Add-ons and updated the Add-ons documentation accordingly.
- Fixed main window frame and shadow rendering on Windows.
- Fixed tray menu focus handling on Windows.

2.8.36 (2026-09-30)

- Moved external Add-ons to the shared application-wide addons directory, with automatic migration from profile workdirs.
- Theme and Locale Add-ons are now loaded directly from the global Add-ons directory without copying files into profiles.
- Improved Add-on localization with private locale directories and dynamic locale domains for plugins, tools and providers.
- Improved Realtime + audio flow, stability and voice interaction handling.
- Added a separate audio input level bar for microphone input.

2.8.35 (2026-09-29)

- Refactored the LiteLLM provider and migrated it to the native LlamaIndex LiteLLM integration. Vision input and tool calling are now supported by compatible LiteLLM models.
- Added an application-wide Package Manager for optional runtime dependencies, with packages stored per Python version under /extra_packages/.
- Added shared text editor preferences for indentation, tab width, and word wrap across the Text Editor, Files preview/editor, and Canvas source editor.
- Added a search field to the Model Importer.
- Added optional model-defined runtime MCP connections, with separate permissions for HTTP/SSE and stdio transports and integration with MCP tool discovery and caching.
- Improved Canvas browser controls, source editing, and persistent address history/search behavior.
- Improved local Whisper dependency installation through the Package Manager.
- Added comprehensive Add-ons examples and API reference documentation. See Documentation -> Add-ons API.
- Added various UI and CSS/QSS fixes and refinements.

2.8.34 (2026-09-28)

- Fixed LiteLLM issues by downgrading to version 1.81.16; fixed issue #214.
- Added line numbers and annotations to file previews in the Files tab.
- Added a recent chats list to the tray menu.
- Added agent final result summaries to system notifications.
- Added support for the .agents directory convention in Agents and Custom Agents.
- Optimized file search in the Files tab.
- Improved annotation handling.
- Improved markers in Notepad.
- Improved CSS/QSS styling.

2.8.33 (2026-09-27)

- Moved old legacy agent modes to Custom Agents; refactored legacy agent workflows; legacy agents now use the same runtime as Agents v2.
- Added column sorting to the Files tool.
- Added search engine support to the Canvas browser.
- Added file preview and editing support to the Files tab.
- Added filters and search input to the Skills, Connectors, and Add-ons lists.
- Added pop-up daily notes to the Calendar tool.
- Added annotations to the chat view via RMB → Annotate for AI.
- Moved the Skills menu to Config.
- Refactored LLM providers; setup options were moved to provider classes.
- Improved toolbox responsiveness.
- UI fixes.

2.8.32 (2026-09-26)

- Added support for external Add-ons (beta), allowing users to extend PyGPT with custom plugins, LLM/providers, vector stores, data loaders, audio input/output providers, web providers, tools, agents, themes, and locale packs. Add-ons can be installed from the public catalog, local files and directories, ZIP archives, and external GitHub repositories.
- Added a public PyGPT Add-ons repository and catalog, where anyone can publish and submit their own Add-ons for discovery and installation directly from PyGPT.
- Improved realtime audio flow: added interruption of the previous response when a new request is sent and fixed auto-follow scrolling.
- Fixed equal split-screen column widths when enabling split-screen mode.
- Agent Workflow tool connected to legacy agents.
- Added auto-updater.
- UI improvements.

2.8.31 (2026-09-25)

- Refactored and improved tab management.
- Refactored and improved message rendering and event flow.
- Improved the Canvas plugin.
- Added the ability to send sketches and images directly from Painter to the model in real time via the Canvas plugin.
- Improved and fixed runtime attachment and image handling.
- Optimized memory usage and Python module initialization.
- Added a new Jev / System One plugin - see the documentation for more information.
- Various UI fixes and improvements.
- Other fixes and improvements.

2.8.30 (2026-09-24)

- Added a new Canvas plugin featuring an interactive, real-time canvas with HTML and JavaScript support. It enables visual prototyping, live annotations, HTML generation, opening and editing websites, a built-in web server, and much more. See the new Canvas section in the documentation for details.
- The chat input field is now pinned to its corresponding chat column.
- Added support for new models: Claude Opus 5.5, GPT-6 Sol, and GPT-6 Luna.

2.8.29 (2026-09-22)

- Renamed Chat with Agents mode to Agents.
- Updated default agent preset models to gpt-5.6-luna.
- Fixed compatibility errors in legacy Agent modes.
- Simplified LlamaIndex fallback for models without native tools to standard LLM.
- Added support for passing a reference image directly from a chat attachment to the image generation tool.
- Added and unified local/remote tool permissions across more agent workflow roles.
- Improved runtime artifact sharing between host, Docker Python/IPython, System/OS tools, and agent workflows.
- Google GenAI SDK upgraded to v2.25.0, OpenAI Agents upgraded to v0.18.3.

2.8.28 (2026-09-22)

- Integrated Chat with Files into the standard Chat mode. RAG is now available across all supported modes.
- Added a Judge mode to Autonomous, providing improved response evaluation and continuation guidance after each step.
- Added support for defining preinstalled packages in the built-in Python Interpreter and System / OS sandboxes.
- Added an option to manually rebuild the built-in sandbox virtual environment.
- Optimized Python command execution.
- Other fixes and improvements.

2.8.27 (2026-09-20)

- Added a built-in sandboxed Python interpreter running in its own virtual environment, managed by uv, available as the default third sandbox option in the Python Interpreter and System / OS plugins. It allows Python code to be executed without requiring Python to be installed on the host system and without using Docker, as the interpreter is bundled and managed directly by PyGPT. See the documentation: Plugins -> Python Interpreter.
- Fixed HTML / Canvas base directory handling for relative paths and local assets.
- UI and CSS fixes, refinements, and visual improvements.

2.8.26 (2026-09-20)

- Refactored, improved, and simplified CSS/QSS handling. Overriding and customizing CSS and QSS is now much easier - see the Extending PyGPT section in the documentation.
- Fixed and optimized RAG retrieval for improved reliability and performance.
- Fixed the splitter resize policy and improved layout resizing behavior.

2.8.25 (2026-09-19)

- Added support for Agent Skills, including importing from GitHub, local files, and formats compatible with Claude, Codex, OpenClaw, and other supported ecosystems. Added a dedicated Skills management interface for browsing, installing, enabling, disabling, and removing skills.
- Added support for Claude/Codex-style Connectors, integrated with the MCP plugin. Connectors can be imported from GitHub, local files, Claude, Codex, OpenClaw, Cursor, VS Code, OpenCode, MCPorter, and compatible JSON, TOML, and YAML definitions. Connector management is available under Config → MCP → Connectors.
- Added new application themes: Matrix, Gray, Mint, Flare, Ocean, Sun, and Retro.
- Added Full Screen mode (F11) and support for a frameless window layout.
- Optimized message sending from the chat input by moving pre-send preparation tasks to asynchronous workers, reducing UI blocking before requests are sent.
- Various UI fixes, layout improvements, workflow fixes, and usability refinements across the application.

2.8.24 (2026-09-18)

- Added Import profile... and Export profile... options to the File menu, allowing complete profiles to be exported and imported, including the database, configuration, files, and application data.
- Added an automatic prompt-injection guard in Settings -> Security -> Auto-prevent prompt injections.
- Moved the model selector to the input field.
- Added a clock to the Calendar.
- Added subdirectory lookup support for @mentions.
- Added runtime switching to the Agent Workflows editor.
- Removed the Edit JSON configs option.
- Improved the chat input field.
- Added a loader to the Docker builder.
- Integrated system notifications with Chat with Agents.
- The input field is now hidden in non-chat tabs.
- Added date and time headers above chat input blocks.
- Added support for referencing other conversations from the database using @mentions with conversation IDs, e.g. What were we talking about in chat @123?
- Added various CSS, layout, and UI improvements.

2.8.23 (2026-09-17)

- Improved and extended Agents workflows.
- Extended and condensed Agents instructions.
- Added the Agent Workflow tool with real-time workflow preview, available as a dialog or tab.
- Fixed auto-scroll and scroll-follow issues.
- Optimized memory usage in the Python interpreter.
- Improved CSS and UI layout.

2.8.22 (2026-09-16)

- Reorganized Context settings into General, Tools, and Advanced handling.
- Added Agent Workflows for Chat with Agents with editable built-in workflows, custom UUID-based profiles, independent prompts, toolbox/Config access, migration support, documentation, and translations.
- Moved Chat with Agents prompt editing from Settings to Agent Workflows.
- Added direct Config -> MCP... access to MCP settings.
- Added Restore tool calls in runtime, independent from database storage and history restore.
- Added Display tool calls JSON in Chats -> Render, with aggregated Tool/Tools status when disabled.
- Improved attachment handling.

2.8.21 (2026-09-16)

- Fixed restoring hidden chats from collapsed columns.
- Added tool call storage settings in Settings -> Context, with Disable, Truncate, and Full modes.
- Added history restore mode settings in Settings -> Agents -> Chat with Agents.
- Added a distinct background highlight for the current day of the week in Calendar.
- Added Undo support to the Clear Image button in Painter.
- Added a System Info dialog to the About menu.
- Added support for custom instructions and system prompts in Chat with Agents under Settings -> Agents -> Chat with Agents -> Advanced.
- Optimized database storage.
- Simplified the token counter display.
- Moved the Stream checkbox to Settings -> Context -> Render.
- Moved the Plain Text mode switch to the icon bar above the input field.
- Various UI fixes and improvements.

2.8.20 (2026-09-15)

- Added an integration layer between Chat with Agents and other conversation modes, allowing conversations to be continued seamlessly when switching between modes.
- Optimized Agents memory storage.

2.8.19 (2026-09-15)

- Split the Plugins list into Popular and Other sections.
- Optimized and simplified the Plugin Settings dialog.
- Fixed status visibility on non-chat tabs.
- Fixed footer state restoration after changing profiles.
- Fixed a Google `show_reasoning` error when used as a background vision model.
- Fixed @mention restoration after application startup.
- Fixed runtime WebView restoration when returning from plain text view on another tab.
- Fixed RAG engine LLM selection in Chat with Agents.
- Added support for runtime attachment and image appending from the workdir.
- Added token counters and usage tracking to Chat with Agents.
- Improved image attachment @mentions.
- Added loaders for heavy operations such as indexing and embedding.
- Updated CSS styles and translations.

2.8.18 (2026-09-15)

- Extended the Chat with Agents workflow and added step-by-step execution mode.
- Fixed context token count display.
- Simplified themes to two built-in options: Dark and Light.
- Added @ mentions in the input field for referencing attachments, files, and directories.
- Added Advanced Context Handling (experimental). Enable it in Settings -> Context. It extends the effective conversation context in real time using a memory cache and automatic context trimming.
- Added support for AGENTS.md in the workdir root. See the documentation for more details.
- Other fixes and improvements.

2.8.17 (2026-09-13)

- Fixed Ollama native tool calls.
- Fixed runtime language switching.
- Fixed tab tooltips.
- Fixed chat reload race conditions.
- Fixed streaming of final responses in Chat with Agents.
- Moved max worker limits to Settings.
- Simplified model configuration.
- Added Text tool to Painter.
- Moved Reasoning Effort to a global switch in the input field.
- Refactored Agents and autonomous mode.
- Updated translations and documentation.
- Removed deprecated options.
- Various UI fixes.

2.8.16 (2026-09-12)

- Experts are now full-featured agents, using the same agent runtime and capabilities as agents in Chat with Agents.
- Experts are now available as tools across all supported modes, allowing them to be invoked directly from anywhere in the application.

2.8.15 (2026-09-11)

- Renamed Agents v2 (beta) mode to Chat with Agents.
- Added 3 separate submodes to Chat with Agents:
- Chat — allows natural conversation with the primary agent, with delegated agents used when needed.
- Orchestrator — the previous Agents v2 behavior, where the primary agent acts only as an orchestrator for other agents.
- Swarm — allows creating and running a dynamically defined group of specialized agents/workers in parallel in the background. Swarm mode can run continuously with a defined number of parallel agents.
- Added the ability to configure a custom working directory per project. Right-click a project in the project list and select a custom workdir.
- Added LLM provider fallbacks to OpenAILike when a model is not yet supported by the native LlamaIndex integrations for OpenAI, Google, Anthropic, or xAI.
- Added keyed memory storage to the Memory plugin.
- Fixed collection of used URLs and attachments in Chat with Files.

2.8.14 (2026-09-10)

- Security, stability, and provider fixes - PR #208 by @atharvaHJoshi.
- Fixed list selectors for vision models by removing checks for the deprecated vision mode.
- Fixed the halt and acknowledgement flow in Computer Use.
- Moved the native Perplexity LlamaIndex provider to the shared OpenAI-compatible wrapper.
- Fixed legacy beta headers for Anthropic in Chat with Files mode.
- Fixed support for remote tools and Computer Use in Chat with Files mode.
- Fixed race conditions and restart loops in the IPython plugin.
- Added splitter anchors to the CSS in the light theme.
- Added a new model: gpt-image-2.5.
- Added an Insert date/time option to the Notepad right-click menu.
- Added support for Completion-only mode with models and providers other than OpenAI `gpt-3.5-instruct`.
- Added support for local Jupyter/IPython in compiled builds.
- Set IPython’s default stdin to DEVNULL when running code with interactive input to prevent freezes.
- Added support for Computer Use in Realtime + Audio, Legacy Agents, and Autonomous modes.
- Various UI and CSS fixes.

2.8.13 (2026-09-09)

- Fixed issue with empty parameters in the Anthropic API remote tool for computer use.
- Added support for the use of computer use remote tool in Agents v2 for Anthropic and Google.
- Added OSINT v2 preset to Agents v2.
- Added a new plugin: Memory (inline).
- Updated IPython Dockerfile: included default installation of pandas, matplotlib, scikit-learn, and other useful libraries.
- Integrated Google remote tool - MCP.
- Added validation for pasting large directory attachments.
- Enhanced CSS in chat view for better aesthetics.
- Improved handling of multiple tabs.
- Fixed issue with hiding date separators in context list view.
- Corrected profile switch components restoration.
- Other fixes.

2.8.12 (2026-09-08)

- Improved, extended, and fixed several bugs in the following modes: Chat, Realtime + Audio, Computer Use, Autonomous Agent, and Agents v2.
- Added new models: GPT-6 Astra and Claude Fable 5.1.
- Added a new remote tool in Settings: Computer Use, which allows you to control the computer in standard Chat mode.
- Added item limits, a Show more option, and collapsible items to the project list.
- Model responses are now split into partials and stored in the database.
- Fixed multi-column context handling.
- Fixed sandbox image URLs.
- Removed redundant tool input/output from message footers.
- Removed old and deprecated models.

2.8.11 (2026-09-06)

- Added 4 predefined presets to Agents v2: Coder, Researcher, Scientist, and Brainstorm.
- Added auto-scroll and a configurable maximum entries limit to the Code Interpreter window.
- Added support for passing sys_exec inputs to the Code Interpreter window.
- Added verbose mode to Agents v2.
- Added tool execution results in the main context for Agents v2 (disabled by default, can be enabled in Settings -> Agents).

2.8.10 (2026-09-06)

- Added Agents v2 (beta), a new advanced orchestrated multi-agent mode with a user-facing Orchestrator and dynamically created specialist worker agents.
- Added asynchronous worker lifecycle management in Agents v2, including create, update, run/reuse, status, wait, stop and remove operations, with concurrent execution for independent workers and runtime-local worker memory.
- Added local plugin tools, provider-side remote tools, RAG index access, shared attachment context, supported native image input, artifact propagation, persistent Orchestrator memory and live workflow status rendering to Agents v2.
- Added native function-calling support with a ReAct compatibility fallback for Agents v2 models where appropriate.
- Agents v2 is currently in beta; its behavior, workflow rules, preset options and provider compatibility may change in future releases.

2.8.9 (2026-09-05)

- Added a new Custom Providers tab to Settings, allowing users to create and use custom API providers at runtime without modifying the source code.
- Added sandboxed system command execution to the Code Interpreter plugin.
- Added an Active checkbox to the uploaded attachments list, allowing attachments to be enabled or disabled at runtime.
- Added rendering of tool results as formatted JSON code blocks.
- Fixed busy-state rendering in Agents workflows.
- Removed image footers from the WebView.

2.8.8 (2026-09-05)

- Fixed tool call responses in Chat with Files when using the Responses API.
- Fixed attachment uploads in Chat with Files mode.
- Fixed an issue where the remote store after_update hook silently failed to run — PR #203.
- Improved agent response rendering.
- Improved vector store cleanup when deleting contexts.
- Added image editing and remixing features to the inline Image Generation plugin, allowing images to be edited and remixed directly in chat.
- Upgraded the default Docker image definitions for sandboxed IPython and System environments: Python 3.12 is now the default, essential packages were added, and containers now run as a non-root user by default.
- Added isolated per-project LlamaIndex indexes, created and resolved automatically as Current project without adding project-specific entries to the regular indexes list.
- Added project-aware context auto-indexing with Off, Auto-index all conversations, and Auto-index only in projects policies, plus a Use isolated index per project option.
- Added incremental project context indexing with per-project progress tracking, project index update/truncate actions, cleanup on project deletion, and index rebuilding when duplicating projects.
- Added project-aware indexing and retrieval to Chat with Files, the Chat with Files (LlamaIndex, inline) plugin, Files I/O, and the Files context menu.
- Added Clear and truncate index management, including permanent deletion of stored indexes and batch truncation of project indexes.
- Improved index/file tracking with lazy loading to avoid loading the complete indexed-files database into memory.
- Reorganized LlamaIndex settings into File indexing, Context indexing, Data loaders, and Clear and truncate sections and clarified that removing an index from the configured list does not delete its stored data.

2.8.7 (2026-09-04)

- Improved Painter with new drawing modes: Free, Arrow, Rectangle, Circle, and Line, making it faster and easier to annotate images and screenshots with common shapes and symbols.
- Improved UI consistency by simplifying selected icons, labels, and controls across the application.
- Fixed and refined UI translations, descriptions, and tooltips for improved clarity and consistency.

2.8.6 (2026-09-03)

- Added tool-chain grouping in the chat view, combining consecutive tool calls into a single expandable group.
- Improved the context list UI: added section headers, added New context / New project actions directly to section rows, added support for collapsing and expanding sections, improved section interaction and layout.
- Renamed Groups to Projects throughout the application.
- Fixed toolbox minimum-width calculation to prevent individual modes or controls from unexpectedly expanding the toolbox.
- Added new UI icons and tooltips for improved usability and consistency.
- Added a visual screen flash / blink effect after screenshot and camera capture.
- Fixed math rendering of < and > symbols in KaTeX expressions.
- Fixed rendering of the
eq /
e math symbols - issue #199.

2.8.5 (2026-09-02)

- Local Whisper audio input improvements: optimized execution and error handling, added an option to keep the model loaded in RAM, and simplified local model configuration.
- Added Vision support for local Ollama models. To enable it, simply enable Image input in the model settings (if the selected model supports vision).
- Fixed Ollama connection in Chat with files mode.
- Added Ask with screenshot → Select region... to the tray menu, allowing interactive selection of a screen region for screenshot capture.
- Added new Ollama models: SpeakLeash/bielik-11b-v3.0-instruct:Q4_K_M, deepseek-r1:8b, gemma4:e4b, llama4:scout, mistral-small3.2, nemotron-3.5-lightning:30b, qwen3.6:27b, and qwen3.5:9b.
- Removed Assistants mode because the OpenAI Assistants API was deprecated and shut down on August 26, 2026.

2.8.4 (2026-08-19)

- Added a new Security settings section for host-side plugin permissions.
- Added separate read and write restrictions for local filesystem access outside the workdir data directory.
- Added per-system Linux, Windows, and macOS system-command whitelists and blacklists, with the whitelist taking precedence when enabled.
- Added shared filesystem and command security checks across filesystem-capable plugins and host-side command execution.
- Added Permission denied tool responses that point to Settings -> Security when filesystem access is blocked.
- Added a Computer Use tab in Security settings.
- Added Halt on potentially unsafe operation, enabled by default.
- Added explicit user confirmation for provider-flagged potentially unsafe Computer Use operations in OpenAI and Google providers.
- Added per-model custom API endpoint and API key settings.
- Improved the Models Editor advanced settings layout and added missing field descriptions and translations.
- Moved application-managed temporary files for audio input, HTML Canvas, Code Interpreter/IPython, and the Transcript tool to the workdir tmp directory.
- Centered the audio input/output level meter in the bottom input bar.
- Fixed the Painter tool restoring the wrong color icon after application restart.
- Fixed escaped Unicode characters in stored and displayed command/tool JSON, including nested tool responses.
- Updated examples and documentation.

2.8.3 (2026-08-16)

- Fixed duplicated attachments being displayed after tool calls.
- Fixed duplicated images in tool chains.
- Fixed attachment paths in additional context so the full filesystem path is no longer sent.
- Fixed URL extraction from responses.
- Fixed reasoning context being appended to responses for local Ollama models such as DeepSeek.
- Added native attachment file upload when supported by the provider.
- Added real-time reasoning rendering.
- Added collapse/expand controls for URLs and attachments lists.
- Added missing settings descriptions.
- Upgraded the image generation plugin to support the newest models.
- Changed attachment archive rendering to display an archive as a single entry.
- Changed tool-chain rendering to display action icons at the end of the chain.
- Set max_total_tokens to 0 by default.
- Updated the context max limit option description.
- Updated documentation and configuration descriptions.

2.8.2 (2026-08-15)

- Added ARM64 AppImage and Snap builds.
- Added support for preserving full archive paths when attaching ZIP files.
- Removed the Send mode: Off option.
- Improved tool call rendering.
- Improved CSS and UI.
- Moved the Clear input on send option to Settings.

2.8.1 (2026-08-13)

- Added LiteLLM as LLM provider - #181, thanks to @RheagalFire.
- Added Forge as LLM provider - #173, thanks to @Yiiii0.
- Added Eden AI as LLM provider - #194, thanks to @MVS-source.
- Added TwelveLabs plugin (video understanding with Pegasus + Marengo embeddings) - #191, thanks to @mohit-twelvelabs.
- Fix: use unique filename per xAI TTS generation to avoid Windows file-lock PermissionError - #192, thanks to LittleBallOfPurr.
- Fix: image and video URLs formatting in Windows.
- Added an option to hide models from the model list in the Models Editor.
- UI improvements.

2.8.0 (2026-08-12)

- Upgraded all models to their latest available versions.
- Upgraded dependencies.
- Fixed compatibility with the latest models.
- Fixed the Runner agents workflow.
- Fixed the Ollama provider and custom model handling - #169, thanks to @ba2512005.
- Fixed top_p handling and added temperature guards for Claude 4.x Anthropic models - #184, thanks to @robertnama.
- Fixed message handling for Mistral - #185, thanks to @WildGreenRose.

2.7.12 (2026-02-06)

- xAI SDK upgraded to v1.6.1.
- Added video generation support in xAI provider.
- Added new models: grok-imagine-image and grok-imagine-video.
- UI improvements and smoother navigation between tabs and columns.
- Optimized adding attachments. Added options to attach an attachment only once (if the API stores messages on its side), and an option to force attaching an attachment only once.
- Updated libraries.

2.7.11 (2026-02-05)

- Improved handling focus between columns.
- Fixed models debugger update.

2.7.10 (2026-02-03)

- Fixed an issue where an avatar could be overwritten when creating a new preset.
- Fixed an issue where a new context was not created when opening a new tab in the second column.
- Added prompt history navigation to the input field (Ctrl + Up/Down arrow keys).
- Added initial image centering when loading the Image Viewer.
- Added a Mark/Unmark feature to the Notepad widget.
- Added 18 new languages: Arabic (ar), Bulgarian (bg), Czech (cs), Danish (da), Finnish (fi), Hebrew (he), Hindi (hi), Hungarian (hu), Japanese (ja), Korean (ko), Dutch (nl), Norwegian (no), Portuguese (pt), Romanian (ro), Russian (ru), Slovak (sk), Swedish (sv), Turkish (tr).

2.7.9 (2026-01-08)

- Improved realtime audio mode.
- Added xAI provider and Grok support in realtime audio mode.

2.7.8 (2026-01-06)

- Added the xAI Collections remote tool and integrated collections management into the Remote Vector Stores tool.
- Unified the Remote Vector Stores tool into a single tool for all providers.
- Added xAI Grok audio input and output provider (realtime audio support for Grok coming soon!).
- Enhanced the Image Viewer tool with a toolbox featuring Prev/Next and more options.

2.7.7 (2026-01-05)

- Added support for Responses API in xAI.
- Added xAI remote tools: Remote MCP, Code Execution.
- Added Anthropic remote tools: Remote MCP, Web Fetch, Code Execution.

2.7.6 (2026-01-03)

- Fixed compatibility with xAI SDK and resolved empty responses from Grok models.
- Fixed missing libraries in the Snap package.
- Added zoom and grab functionality in the Image Viewer.
- Added a zoom menu to textarea and web widgets.
- Added the ability to close tabs with a middle mouse button click.

2.7.5 (2026-01-03)

- Added Sandbox/Playwright option to Computer Use mode.
- Added support for Google models in Computer Use mode and introduced a new model: gemini-2.5-computer-use-preview-10-2025.
- Added support for Google models in Research mode and introduced a new model: deep-research-pro-preview-12-2025.
- Added the Google Vector Stores tool.

2.7.4 (2025-12-31)

- Added a splash screen.
- Added Preview and Download links to the Image and Video outputs.
- Added Negative prompt input to Image and Video mode.
- Improved focus handling.
- UI improvements.

2.7.3 (2025-12-30)

- Added the `Remix/Extend` option in Image and Video generation mode. This allows the use of a previously generated image or video as a reference. It can be used for adding or changing elements in a previously generated image or video instead of creating a new one from scratch. See the docs: `Modes -> Image and Video generation -> Remix, Edit, or Extend`.

2.7.2 (2025-12-29)

- Fixed: non-searchable combobox width.
- Improved updater.
- Added .AppImage build.

2.7.1 (2025-12-28)

- Improved UI elements.
- Optimized Painter rendering and redraw functions.
- Added Pack/Unpack feature to File Explorer.
- Fixed: image restoration in Painter.
- Fixed: tab title updating upon context deletion.

2.7.0 (2025-12-28)

- Added multi-select functionality using CTRL or SHIFT and batch actions to the context list, preset list, attachments list, and other list-based widgets.
- Added a search field to comboboxes, such as the model selector.
- Added a Duplicate option to the models editor.
- Added drag-and-drop to context list.
- Added multi-select, drag-and-drop, Cut, Copy, and Paste features to the File Explorer.
- Fix: scroll restoration after actions in the context list.
- Fix: 'Use as image' option in the File Explorer.
- Fix: current preset system prompt disappearing on profile change.
- Other UI fixes/improvements.

2.6.67 (2025-12-26)

- Added a provider filter to the models editor.
- Added video options (resolution, duration) to the toolbox.
- Updated the models configuration.

2.6.66 (2025-12-25)

- Added Sora 2 support - #155.
- Added Nano Banana support.
- Added Qdrant Vector Store - merged PR #147 by @Anush008.
- Added models: gpt-5.2, gpt-image-1.5, gemini-3, nano-banana-pro, sora-2, claude-sonnet-4.5, claude-opus-4.5, veo-3.1.
- Added Select/unselect All option in checkbox lists.
- OpenAI SDK upgraded to 2.14.0, Anthropic SDK upgraded to 0.75.0, xAI SDK upgraded to 1.5.0, Google GenAI upgraded to 1.56.0, LlamaIndex upgraded to 0.14.10.
- Fix: charset-normalizer 3.2.0 circular import - #152.
- Fix: Google client closed state.

2.6.65 (2025-09-28)

- Added drag and drop functionality for files and directories from the filesystem in attachments and file explorer.
- Added automatic thumbnail generation when uploading avatars.
- Added a last status timer.
- Added a fade effect to collapsed user messages.
- Added a scroll area to the agent options in the presets editor.
- Added a hover effect to lists.
- Improved UI/UX.

2.6.64 (2025-09-27)

- Added translations to agent headers.
- Improved presets tabs.
- Added support for music (Lyria) in both image and video modes (beta).

2.6.63 (2025-09-27)

- Improved agents' workflows.
- Enhanced the display of agents' steps in the UI.

2.6.62 (2025-09-26)

- Enhanced agent workflow execution.
- Improved preset list handling by adding a drop field indicator and fixing auto-scroll.
- Added middle-mouse button panning to Painter.
- Added an input character counter.

2.6.61 (2025-09-26)

- Enhanced the agents node editor, custom agent flow, and instruction following.
- Added drag-and-drop and reordering functionality to the presets list.
- Added statistics for response tokens, including time elapsed and tokens per second.
- Improved UI/UX.

2.6.60 (2025-09-25)

- Added a new tool: Agents Builder - allowing visual design of agent workflows using nodes - available in Tools -> Agents Builder (beta).

2.6.59 (2025-09-23)

- LlamaIndex has been upgraded to v0.13.6.
- The Responses API is disabled by default in Agents (LlamaIndex) for compatibility.
- Added WinAPI support to the System (OS) plugin.

2.6.58 (2025-09-22)

- Added: internal Web Browser tool.

2.6.57 (2025-09-22)

- Changed: The web search icon has been moved to the input area.
- Improved: The local web search plugin has been enhanced to retrieve multiple URLs at once.
- Added: Use proxy switch in Settings.

2.6.56 (2025-09-22)

- Optimized: Memory usage and performance in streaming and rendering large contexts.
- Added: Copy to clipboard functionality in user messages.
- Added: Manual scroll in plain-text mode.

2.6.55 (2025-09-18)

- Fixed: Unnecessary context loading from the database.
- Optimized: Token count in the input field.

2.6.54 (2025-09-18)

- Added: Remote tools (like web search) are now also available in the Chat with Files and Agents (LlamaIndex) modes.
- Added: Two new plugins: Wolfram Alpha and OpenStreetMap.
- Fixed: Enabled local file-like schemes in links/images in the markdown-it parser.

2.6.53 (2025-09-17)

- Added: An icon to enable/disable the web search remote tool in the icon bar, along with remote web search functionality in OpenRouter (#135).
- Added: The ability to mute audio in real-time mode via the audio icon.

2.6.52 (2025-09-17)

- Added MCP plugin: Provides access to remote tools via the Model Context Protocol (MCP), including stdio, SSE, and Streamable HTTP transports, with per-server allow/deny filtering, Authorization header support, and a tools cache.
- Fixed: tab tooltips reload on profile switch.

2.6.51 (2025-09-16)

- Fix: Automatically reloading calendar notes.
- Fix: Context menu CSS background color for calendar items.

2.6.50 (2025-09-16)

- Optimized: Improved memory cleanup when switching profiles and unloading tabs.
- Fix: Resolved missing PID data in text output.
- Fix: Enhanced real-time parsing of execute tags.

2.6.49 (2025-09-16)

- Fixed: Occasional crashes when focusing on an output container unloaded from memory in the second column.

2.6.48 (2025-09-15)

- Added: auto-loading of next items to the list of contexts when scrolling to the end of the list.

2.6.47 (2025-09-15)

- Improved: Parsing of custom markup tags.
- Optimized: Switching profiles.

2.6.46 (2025-09-15)

- Added: Global proxy settings for all API SDKs.
- Fixed: xAI client configuration in Chat with Files.
- Fixed: Top margin in streaming container.
- Refactored: Debug workers.

2.6.45 (2025-09-13)

- Improved: Parsing of custom markup in the stream.
- Improved: Message block parsing moved to JavaScript.

2.6.44 (2025-09-12)

- Added: Auto-collapse for large user input blocks.
- Added: Configuration for syntax highlighting intervals.
- Improved: Visibility of label icons.
- Improved: Scrolling of code blocks.
- Fixed: Parsing of quotes in custom markdown blocks.

2.6.43 (2025-09-12)

- Fixed: preset restoration when switching profiles.
- Improved: faster application launch and exit.

2.6.42 (2025-09-11)

- Fixed: Save/load zoom factor in the chat window when switched via Ctrl + mouse wheel.
- Fixed: Break the first line in code blocks.
- Added: Configuration options for syntax highlight limits.
- Updated: SVG icons.

2.6.41 (2025-09-11)

- Rendering engine optimizations: markdown parsing moved to JS, reduced CPU usage, added auto-memory clearing, and more.

2.6.40 (2025-09-06)

- CSS fixes.

2.6.39 (2025-09-06)

- Added: Search input to the models editor.
- Fixed: Brush color switching when changing modes in the Painter.

2.6.38 (2025-09-05)

- Fixed: Detection of chunk type in Ollama.
- Fixed: Import of models with existing IDs.
- Fixed: Updating the Assistants UI list after creating a new Assistant.
- Refactored and optimized the code.


2.6.37 (2025-09-05)

- Fixed: Function parameters sanitization in the Google Gen AI SDK.
- Added: Native support for Anthropic SDK and xAI SDK (with remote tools).

2.6.36 (2025-09-04)

- Optimized rendering of large code blocks.

2.6.35 (2025-09-04)

- Improved responsiveness in chat streaming.
- Added zoom, fit, and auto-scroll on crop in Painter.
- Added a scroll-up button in the chat window.
- Optimized memory usage.

2.6.34 (2025-09-03)

- Added auto-resizing for input.
- Added file management to the OpenAI Vector Stores tool.
- Implemented removal of script tags when exporting HTML.
- Enabled maintaining custom resolution in Painter.

2.6.33 (2025-09-02)

- Added a "crop" option, auto-resize canvas, and layer support to Painter.
- Improved video capture from the camera.

2.6.32 (2025-09-02)

- Added video generation and support for Google Veo 3 models.
- Introduced new predefined models: veo-3.0-generate-preview and veo-3.0-fast-generate-preview.
- Integrated DuckDuckGo as a search provider in the WebSearch plugin.
- Added "Loop" mode to Realtime + audio mode for automatic turn handling and continuous conversation without manually enabling the microphone.

2.6.31 (2025-09-01)

- Chat with Audio mode renamed to Realtime + audio.
- Added support for real-time audio models from OpenAI (Realtime API) and Google (Live API), featuring real-time audio integration (beta).
- Introduced new predefined models: gpt-realtime, gpt-4o-realtime-preview, and gemini-2.5-flash-preview-native-audio-dialog.
- Included Google Gen AI audio input and output providers in the Audio Input/Output plugins.
- Added URL Context remote tool support in Google Gen AI.

2.6.30 (2025-08-29)

- Added native Google GenAI API support (beta); live audio is not supported yet (#132).
- Added new predefined models for image generation: Google Imagen3 and Imagen4.
- Optimized token usage in the Responses API.
- Removed Vision mode (it is now integrated into Chat).

2.6.29 (2025-08-28)

- Verbose options have been moved to the Developer section in settings.
- Enhanced logging of embeddings usage.
- Fixed styles list.

2.6.28 (2025-08-27)

- Added new plugins: Tuya (IoT) and Wikipedia.
- Improved formatting of JSON command output.
- Fixed CSS issues.

2.6.27 (2025-08-26)

- Simplified audio input: A microphone icon has been added to the input field.

2.6.26 (2025-08-26)

- Added new provider: OpenRouter (#133).

2.6.25 (2025-08-26)

- Fixed the empty agent ID issue in OpenAI Agents evaluation.
- Added the ability to select a custom model for evaluation.
- Added embedding providers: Anthropic, Deepseek, MistralAI, xAI, VoyageAI.

2.6.24 (2025-08-26)

- Added a new option: LlamaIndex -> Embeddings -> Default embedding providers for attachments.
- The same model provider is now used for both embedding and RAG query in attachment indexing.
- Translations have been added to Agents.
- Fixed fetching Anthropic models list.
- Added Google GenAI Embeddings.

2.6.23 (2025-08-25)

- Added an inline "Add a new chat" button to the right of the tabs.
- Added an "Add Attachment" button in the input field.
- Improved file open in the system's file manager
- Fixed the restoration of input text color when changing themes from light to dark.
- Fixed last eval step finish if 100% complete.

2.6.22 (2025-08-25)

- UI refactor and optimizations.

2.6.21 (2025-08-24)

- Ollama models are now available in OpenAI Agents mode.
- Improved parsing of responses from Agents.
- Fix: do not initialize index in Agents mode if not provided.
- Fix: agent response evaluation steps limit.
- Fix: do not execute code in agents if Tools are disabled.
- Refactoring.

2.6.20 (2025-08-22)

- Added a new plugin: Server (FTP/SSH) - connect to remote servers using FTP, SFTP, and SSH. Execute remote commands, upload, download, and more (beta).
- Added support for Wayland in Snap/compiled versions.

2.6.19 (2025-08-22)

- Fixed: added prevention for summarizing an empty context.
- Improved the speed of context item refreshing.

2.6.18 (2025-08-21)

- Refactor and optimizations.
- Fix: Evolve agent stop event calling.

2.6.17 (2025-08-21)

- Optimized profile switching.
- Fixed: setting initial splitter size on first launch.
- Added smoother view reload.

2.6.16 (2025-08-20)

- Fixed: Attachment string joining.
- Improved expert function calls.
- Added sorting to the plugin list.

2.6.15 (2025-08-20)

- Added: do not change the context menu font size in text editing.
- Added: do not reload context items on tab change if already loaded.
- Fixed: appending of names and avatars in the stream chunk.

2.6.14 (2025-08-19)

- Fixed: Agent evaluation tool runs even if tools are disabled.
- Extended agent response evaluation by providing the full context of the output.

2.6.13 (2025-08-19)

- Fix: Do not load the index in experts if it is not provided.
- Fix: Load remote images in the webview.
- Fix: Presets list refresh.
- Optimize context items reload.

2.6.12 (2025-08-19)

- Optimized web renderer memory cleanup.

2.6.11 (2025-08-18)

- Added the ability to close the dialog window with the Esc key.
- Made context item deletion without output refresh.
- Optimizations.

2.6.10 (2025-08-17)

- Enhanced the handling of the context list.
- Integrated RAG into OpenAI Agents.
- Enhanced RAG management in Agents.
- Added an option: Config -> Agents -> General -> Auto-retrieve additional context from RAG.
- Included Google Docs, Maps, and Colab in the Google plugin.

2.6.9 (2025-08-17)

- Added two new agents for LlamaIndex and OpenAI: Supervisor and Worker (beta).

2.6.8 (2025-08-16)

- Fixed: updated paragraph color on theme switch.
- Added switching to duplicated preset after creation.
- Reduced delay after selecting context.
- Optimized rendering of mathematical formulas.

2.6.7 (2025-08-16)

- Fix: missing entity sanitize.

2.6.6 (2025-08-16)

- Output rendering optimization.

2.6.5 (2025-08-16)

- Fix: crash when creating a context in a new group.
- Fix: high CPU usage after switching profiles.

2.6.4 (2025-08-15)

- Fix: tool execution in OpenAI Agents.
- Optimizations.

2.6.3 (2025-08-15)

- Optimized streaming and CPU usage.
- Fixed crash on set label color and ctx duplicate.

2.6.2 (2025-08-15)

- Added plugins (beta): Google, Facebook, X/Twitter, Telegram, Slack, GitHub, Bitbucket.

2.6.1 (2025-08-14)

- LlamaIndex Agents refactored to Workflows.

2.6.0 (2025-08-13)

- Added split responses to the OpenAI Agents in non-streaming mode.
- Disabled auto-scroll when manually scrolled to the top.
- Increased scrollbar width in the light theme.
- Optimized the clearing of the streaming buffer.
- Optimized imports.
- Made CSS improvements.

2.5.98 (2025-08-12)

- Added support for GPT-5 in LlamaIndex/Chat with Files mode.
- Experts are now allowed in all OpenAI agent types.
- Improved the output of OpenAI agents (separated context items).
- Refactored memory cleanup for thread workers.
- Optimized streaming.

2.5.97 (2025-08-11)

- Fix: attribute error in prev ctx.

2.5.96 (2025-08-10)

- Fixed memory leaks.

2.5.95 (2025-08-09)

- Added user info personalization in Config -> Personalization, where you can provide information about yourself to the model.
- Added presets personalization with configurable AI names and avatars.
- Added a search field in the Translator tool.
- Fixed <> tags replacement in code blocks.

2.5.94 (2025-08-09)

- Added a new LLM provider: HuggingFace Router.
- Introduced a new model: gpt-oss (OpenAI open-source model available in HuggingFace and Ollama).
- Added a new agent mode in OpenAI Agents: Bot 2 Bot.
- Fixed: Storing the last used context ID when empty.
- Fixed: Reloading items when an agent run is stopped.

2.5.93 (2025-08-08)

- Added a new tool: Translate - in menu Tools - feature #123.

2.5.92 (2025-08-08)

- Added max files to store config option in Audio -> Cache.

2.5.91 (2025-08-08)

- Added GPT-5.
- Added audio cache - #118.

2.5.90 (2025-08-07)

- Fix: Initialize context summary if a conversation starts with a tool call.
- Fix: Store splitter positions even if the object is deleted from memory.
- Update: CSS improvements.

2.5.89 (2025-08-07)

- Added audio output device selection in Config -> Audio - issue #117
- Added audio input and output backend selections in Config -> Audio.

2.5.88 (2025-08-06)

- Optimized the process of unloading tabs from memory.
- Reduced initial RAM usage on launch.
- Added a handler for the SIGTERM signal.

2.5.87 (2025-08-05)

- Optimized memory cleanup.

2.5.86 (2025-08-04)

- Optimized CPU and memory usage.
- OpenAI vector stores tool added to Tools menu.
- Fixed schema parsing for tools in Agent (OpenAI) mode.
- Fixed multi-threading when uploading to remote vector store.
- Fixed Urls open in Snap.

2.5.85 (2025-08-02)

- Added importer for models from providers in Config -> Models -> Import... - #127
- Fix: options save in models editor.

2.5.84 (2025-08-02)

- Added a new OpenAI agent mode: Evolve. You can find its description in the documentation under the section Modes -> Agent (OpenAI).

2.5.83 (2025-08-01)

- Improved streaming in Agent (OpenAI) mode.
- Improved loading of default options in presets.
- Added context summary event if the kernel stopped.
- Implemented dynamic width for opened combo boxes.

2.5.82 (2025-08-01)

- Added a new OpenAI agents: researcher, planner, feedback.

2.5.81 (2025-07-31)

- Disabled remote tools by default in non-OpenAI providers in Agent (OpenAI) mode.
- Removed unsupported models from Agent (OpenAI) mode.

2.5.80 (2025-07-30)

- Improved the stop command in Agent mode: added saving of current output on break.
- Fixed the experts providers list in Agent (LlamaIndex): now only Llama providers are displayed.
- Added max_turns configuration to Agent (OpenAI), shared with Agents -> Max steps configuration option value.

2.5.79 (2025-07-30)

- Added prevent text break on input send when the cursor is inside the text.
- Fix: updating expert IDs on preset save.
- Added notification if no agent is selected.
- Disabled FFmpeg warnings.

2.5.78 (2025-07-30)

- Added support for Python 3.13.
- Fixed history prepare in Agent (LlamaIndex) mode.

2.5.77 (2025-07-30)

- Fix: history prepare in Agent (OpenAI) mode.
- Agent response evaluation splitted into two modes: by the percentage of task completion and by accuracy.

2.5.76 (2025-07-30)

- Added a new mode: Agent (OpenAI) and integrated `openai-agents` into the app (beta).

2.5.75 (2025-07-28)

- Fix: context append in LlamaIndex agents.

2.5.74 (2025-07-28)

- Fix: mouse buttons handling in Computer use mode.
- Added double click handler.

2.5.73 (2025-07-28)

- Stream rendering optimization, reduced CPU usage.
- Fix: < and > rendering in math formulas.
- Fix: do not automatically add tools to the agent when the inline plugin is enabled but the Tools option is not enabled.
- Fix: corrected saving of the last context element when the stream is interrupted.
- Fix: unlock response regeneration after stopped event.

2.5.72 (2025-07-27)

- Improved stop command.

2.5.71 (2025-07-27)

- Added a new working mode: Computer Use for autonomous navigation in the user's environment (beta; utilizes the Computer Use remote tool and the model computer-use-preview).
- Added a new remote tool: Remote MCP (with configuration in Settings -> Remote Tools).
- Added a new remote tool: File Search (with configuration in Settings -> Remote Tools).
- Added a new option: Always continue... to Agent (autonomous) plugin settings.

2.5.70 (2025-07-26)

- Added separate config for Responses API for expert instances.
- Added a new model: mistral-small3.1.

2.5.69 (2025-07-25)

- The Responses API and remote tools are now allowed in Agent (autonomous) and Expert modes. Default: disabled.
- Added separate options in the configuration for enabling the Responses API in: Config -> Agents and Experts.
- Improved expert and agents system prompt.

2.5.68 (2025-07-25)

- Added a separate configuration to enable or disable native function calls in both agent and expert modes.

2.5.67 (2025-07-25)

- Added native tool call functionality in Experts.
- Enhanced the use of multiple tool calls in Experts.
- Fixed the display of Expert names when responding.
- Improved context history handling in Experts.

2.5.66 (2025-07-24)

- Fixed the issue where input focus was lost on the initial context summary.
- Added input focus at application startup.

2.5.65 (2025-07-24)

- Improved expert and agent calls and response rendering.
- Fixed duplicate goal completion in autonomous mode.

2.5.64 (2025-07-23)

- Updated the system prompts.
- Allowed native tool calls in Chat with Files when the agent is not enabled.

2.5.63 (2025-07-23)

- Agent mode improvements.
- Custom system prompt allowed for all agent types.

2.5.62 (2025-07-23)

- Enhanced system prompt append to planner agent in Experts.

2.5.61 (2025-07-23)

- Enhanced expert and autonomous agent modes.
- Added streaming output from the agent in Chat with Files mode.
- Introduced a planner agent as the engine for experts.
- Fixed issue with goal completion in autonomous agent mode.
- Resolved CSS issues with sliders on Windows.

2.5.60 (2025-07-22)

- Improved CSS.
- Added a new color themes: Light, Light:Gray, Dark, Dark:Gray, Dark:Darker.

2.5.59 (2025-07-22)

- Added missing locales.
- Fix tab title rename on context rename.

2.5.58 (2025-07-21)

- Fix: Google API tool prepare requests.
- Added delay to search input update.
- Added per provider empty API key monits.

2.5.57 (2025-07-21)

- Fix: context summary worker initialization.
- Fix: loading spinner in Assistants mode.

2.5.56 (2025-07-21)

- Improved data reload after profile switch.

2.5.55 (2025-07-21)

- Fixed tab title reload.
- Fixed restore selection after pin/unpin.
- Added tabs debugger.
- Removed JSON output string from tool output.

2.5.54 (2025-07-20)

- Fixed reloading chat items if second column is focused.
- Increased chat tab titles to 15 chars.
- Removed gpt-4.5-preview from models - issue #121.
- Updated CSS for images.

2.5.53 (2025-07-19)

- Improved context list context menu.
- Improved presets list context menu.
- Chat icons changed to SVG.
- API organization key added to container file download request.
- Fixed switch to idle after tool call in agent mode.

2.5.52 (2025-07-19)

- Added automatic files and images download from Response API / remote Code Interpreter containers.

2.5.51 (2025-07-18)

- Updated CSS, removed uppercase from tabs and buttons.
- Fixed tab names translations.

2.5.50 (2025-07-18)

- Improved tabs handling in split-screen.

2.5.49 (2025-07-18)

- Added binary image output handling to docker kernel in Code Interpreter.
- Added score display in agent loop/evaluation mode.
- Added auto-recovery from base config if config file is corrupted.
- Profile move and duplicate moved to separated thread.

2.5.48 (2025-07-17)

- Added binary image output handling from Code Interpreter.
- Added auto-restart IPython kernel on failure or not initialized.
- Improved agents loop and evaluation.
- Added LlamaIndex provider config for Grok.

2.5.47 (2025-07-17)

- Disabled output clear on kernel restart in Code Interpreter.

2.5.46 (2025-07-17)

- Fix: kernel resume after Docker image rebuild.
- Fix: update history after output clear in Code Interpreter.

2.5.45 (2025-07-16)

- Improved IPython kernel restart in Docker container.
- Fixed Ctrl+F Find shortcut.
- Decreased tabulation size in text inputs.
- Added kernel restart and clear output to Code Interpreter RMB context menu.

2.5.44 (2025-07-16)

- Code syntax highlighting added to the Code Interpreter tool.
- Image output functionality integrated into the Code Interpreter tool.
- Improved/fixed connection with Docker containers.
- Fixed parsing of integer parameters from plugins.
- Fixed 'Add tool' option in tabs + button.

2.5.43 (2025-07-15)

- Improved IPython code execution.
- Added lock for kernel restarting in Code Interepreter to prevent crash if kernel not ready.

2.5.42 (2025-07-15)

- Connected Run code button in python code blocks to IPython (if enabled).
- Updated CSS.

2.5.41 (2025-07-14)

- Fixed: vector stores handling in Assistants API SDK.
- Added: Run button to Python code blocks for real-time code executing in Code Interpreter.
- Updated documentation.

2.5.40 (2025-07-14)

- Enhanced system prompts.
- Increased default Ollama request timeout to 300s.
- Improved function calling.
- Added option 'Tool calls (native)' to models config.

2.5.39 (2025-07-13)

- Improved multiple notepads handling.

2.5.38 (2025-07-13)

- Fixed: empty models list in presets editor.
- Fixed: focus to second column on open Code Interpreter dialog.
- Added: agent providers names.
- Added: ReAct (Workflow) agent.
- Improved: split-screen switch.

2.5.37 (2025-07-13)

- Improved multi-threading and async calls.
- Added loading spinner.
- Added live output from CodeAct agent.
- UI fixes.

2.5.36 (2025-07-12)

- Improved UI, updated CSS.
- Added: collapse/expand button to code blocks.
- Added: preview code button to HTML code blocks.
- Added: provider separators in models list.
- Added: grok-4 model.

2.5.35 (2025-07-11)

- Improved CodeAct Agent: extended prompt, added work directory handling, auto-installation of missing libraries, and generation of images and plots.

2.5.34 (2025-07-11)

- Added: tool calls in CodeAct Agent.
- Added: multimodal/vision in Chat with Files mode.

2.5.33 (2025-07-10)

- Added: CodeAct Agent, integrated with Code Interpreter plugin (beta).
- Fixed: IPython kernel restarting in Code Interpreter plugin.

2.5.32 (2025-07-10)

- Fixed: File manager opening in Snap version.
- Fixed: Maximum scroll position calculation in the notepad.
- Fixed: Updating the message list after regenerating/editing a response.
- Added: Option to hide timestamps in the web view.

2.5.31 (2025-07-10)

- Added: Mistral AI provider - feature #99.
- Added: A new option in Config -> Settings -> Indexes (LlamaIndex) -> Chat -> Auto-retrieve additional context.
- Fixed: Voice control using local Whisper model empty context - issue #113.

2.5.30 (2025-07-09)

- Fixed: Passing config to placeholders in LlamaIndex model kwargs - issue #112.
- Fixed: Custom commands plugin unhashable dict bug - issue #108.
- Fixed: Attempt to read .env in the /home directory in Snap - issue #116.
- Downgraded Numpy to < 2.0.

2.5.29 (2025-07-09)

- Fixed: context auto-summary if history is disable.
- Added: Anthropic base provider.
- Added: multimodal config in models.
- Added: vision in non-GPT models in Chat mode.

2.5.28 (2025-07-08)

- Fixed: Restoring the scrollbar value in the notepad.
- Fixed: Improved math formula rendering.
- Fixed: "Save as Text" option functionality.
- Added: Set the model key to equal the model ID when saving in the model editor.

2.5.27 (2025-07-01)

- Added Code Interpreter tool to Responses API. Disabled by default, you can enable in: Settings -> Remote Tools.
- Added a new models: o3-deep-research, o4-mini-deep-research, codex-mini-latest.
- Added a new option in Settings -> Indexes: Use ReAct agent for Tool calls in Chat with Files mode.

2.5.26 (2025-07-01)

- Improved mode switch in background.

2.5.25 (2025-07-01)

- Tool calls in Chat with Files mode moved to ReAct agent.

2.5.24 (2025-06-30)

- Added attachments and image edits in Image mode.
- Extended tool calls in LlamaIndex mode.
- Added llama-index-llms-google-genai to dependencies.

2.5.23 (2025-06-30)

- Perplexity added to LlamaIndex providers.
- OpenAI Responses API support (with built-in tools) added to LlamaIndex and Agent modes.
- Improved models and indexes configuration in plugins.

2.5.22 (2025-06-29)

- Improved models editor.

2.5.21 (2025-06-28)

- Fixed JS errors in logger.
- Updated CSS.

2.5.20 (2025-06-28)

- LlamaIndex upgraded to 0.12.44.
- Langchain removed from the list of modes and dependencies.
- Improved tools execution.
- Simplified model configuration.
- Added endpoint configuration for non-OpenAI APIs.

2.5.19 (2025-06-27)

- Added option to enable/disable `Responses API` in `Config -> Settings -> API Keys -> OpenAI`.
- Added support for xAI / Grok models, added grok-3 models.

2.5.18 (2025-06-26)

- Non-GPT models are now available in standard Chat mode.
- Added a new remote tool: `image_generation` in Responses API -> disabled by default, enable in `Config -> Settings -> Remote Tools`. Enables native image generation and editing of uploaded images in Chat mode.
- Added a new model `gpt-image-1` and improved image generation.
- Other small fixes.

2.5.17 (2025-06-25)

- Added settings for enable/disable Remote Tools via Responses API in Chat mode: Config -> Settings -> Remote tools. Currently only web-search-preview tool is available, rest of tools coming soon.
- Fixed context summarization in Ollama provider.

2.5.16 (2025-06-25)

- OpenAI API upgraded to 1.91.0.
- Chat mode migrated to Responses API with native built-in web search tool. (beta)
- Fixed file_read tool in I/O plugin.

2.5.15 (2025-06-24)

- Added Ollama models importer in "Config -> Models -> Import from Ollama".
- Fixed Ollama provider in the newest LlamaIndex.
- Added the ability to set a custom base URL for Ollama -> ENV: OLLAMA_API_BASE.

2.5.14 (2025-06-23)

- Fix: crash if empty shortcuts in config.
- Fix: UUID serialization.

2.5.13 (2025-06-22)

- Disabled auto-switch to vision mode in Painter.
- UI fixes.

2.5.12 (2025-06-22)

- Fixed send-mode radio buttons switch.
- Added a new models: qwen2.5-coder, OpenAI gpt-4.1-mini.

2.5.11 (2025-06-21)

- Added a new models: OpenAI o1-pro and o3-pro, Anthropic Claude 4.0 Opus and Claude 4.0 Sonnet, Alibaba Qwen and Qwen2.
- Updated Bielik model from v2.2 to v2.3 / merged PR #101.
- Fixed HTML output formatting.
- Fixed empty index in chat mode.

2.5.10 (2025-03-06)

- Added a new model: Claude 3.7 Sonnet.
- Fixed the context switch issue when the column changed and the tab is not a chat tab.
- LlamaIndex upgraded to 0.12.22.
- LlamaIndex LLMs upgraded to recent versions.

2.5.9 (2025-03-05)

- Improved formatting of HTML code in the output.
- Disabled automatic indentation parsing as code blocks.
- Disabled automatic scrolling of the notepad when opening a tab.

2.5.8 (2025-03-02)

- Added a new mode: Research (Perplexity) powered by: https://perplexity.ai - beta.
- Added Perplexity models: sonar, sonar-pro, sonar-deep-research, sonar-reasoning, sonar-reasoning-pro, r1-1776.
- Added a new OpenAI model: gpt-4.5-preview.

2.5.7 (2025-02-26)

- Stream mode has been enabled in o1 models.
- CSS styling for tags (reasoning models) has been added.
- The search input has been moved to the top.
- The ChatGPT-based style is now set as default.
- Fix: Display of max tokens in models with a context window greater than 128k.

2.5.6 (2025-02-03)

- Fix: disabled index initialization if embedding provider is OpenAI and no API KEY is provided.
- Fix: embedding provider initialization on empty index.

2.5.5 (2025-02-02)

- Fix: system prompt apply.
- Added calendar live update on tab change.
- Added API Key monit at launch displayed only once.

2.5.4 (2025-02-02)

- Added new models: `o3-mini` and `gpt-4o-mini-audio-preview`.
- Enabled tool calls in Chat with Audio mode.
- Added a check to verify if Ollama is running and if the model is available.

2.5.3 (2025-02-01)

- Fix: Snap permission denied bug.
- Fix: column focus on tab change.
- Datetime separators in groups moved to right side.

2.5.2 (2025-02-01)

- Fix: spinner update after inline image generation.
- Added Ollama suffix to Ollama-models in models list.

2.5.1 (2025-02-01)

- PySide6 upgraded to 6.6.2.
- Disabled Transformers startup warnings.

2.5.0 (2025-01-31)

- Added provider for DeepSeek (in Chat with Files mode, beta).
- Added new models: OpenAI o1, Llama 3.3, DeepSeek V3 and R1 (API + local, with Ollama).
- Added tool calls for OpenAI o1.
- Added native vision for OpenAI o1.
- Fix: tool calls in Ollama provider.
- Fix: error handling in stream mode.
- Fix: added check for active plugin tools before tool call.

2.4.57 (2025-01-19)

- Logging fix.

2.4.56 (2025-01-19)

- Improved tab switching and focus change.
- Improved global keyboard shortcuts handling.

2.4.55 (2025-01-18)

- Added a new option in settings: Audio -> Recording timeout.
- Added a new option in settings: Audio -> Enable timeout in continuous mode.

2.4.54 (2025-01-18)

- Audio output switched from PyGame to PyAudio. It may be necessary to manually connect Alsa in Snap version with: "sudo snap connect pygpt:alsa".
- Added audio output volume progress bar.

2.4.53 (2025-01-17)

- Fix: issue #89

2.4.52 (2025-01-17)

- Improved audio input button visibility toggle.
- Fix: check for required arguments - issue #88.
- UI Fixes.

2.4.51 (2025-01-17)

- Added a "Continuous recording" mode under Audio Input in the Notepad tab, allowing for recording long voice notes and real-time auto-transcription. (beta)
- A new option has been added in Settings -> Audio -> Continuous recording auto-transcribe interval.

2.4.50 (2025-01-16)

- Refactored audio input core.
- Added audio input volume progress bar.

2.4.49 (2025-01-16)

- Fix: stream render in Assistants mode.
- Fix: items remove in context regen/edit.

2.4.48 (2025-01-16)

- Fix: parsing lists in data loaders configuration.
- Fix: crash on Windows on PySide6 v6.6.0.
- Added Gemini embeddings to LlamaIndex settings.
- LlamaIndex upgraded to 0.12.11.
- Security updates.

2.4.47 (2025-01-14)

- Added support for Python 3.12.
- Added a new model to Chat with Files: gemini-2.0-flash-exp.
- PySide6 upgraded to 6.6.0.

2.4.46 (2024-12-16)

- Added a new tab in Settings: "API Keys", where the API keys configuration for Google and Anthropic models has been relocated.
- Introduced a new mode in "Chat with Files": "Retrieve Only", which allows for retrieving raw documents from the index.
- Fixed a bug related to tool calls in the Gemini provider when using Chat with Files mode.

2.4.45 (2024-12-16)

- Enhanced web data loaders UI.

2.4.44 (2024-12-16)

- Enhanced web data loaders.
- Web loaders have been added to attachments, allowing external web content to be attached to context via the "+Web" button in the Attachments tab.
- Improved handling of attachments in groups and added an attachment icon when a group contains attachments.

2.4.43 (2024-12-15)

- Fix: Bug on attachment upload.
- Added: Attachments uploaded in groups are now available for all contexts in the group (beta).

2.4.42 (2024-12-15)

- Added Mailer plugin, which allows sending and retrieving emails from the server, and reading them. It currently supports only SMTP.
- Added 'web_request' command to the Web Search plugin, enabling GET/POST/PUT and other connections to any address and API endpoint. It also supports sending POST data, files, headers, cookies, and more.
- Improved audio output.
- Enhanced visibility of the Video menu.
- Other fixes.

2.4.41 (2024-12-14)

- Improved switching between columns on a split screen.
- Added visual identification of the active column.

2.4.40 (2024-12-13)

- Enhanced Split Screen mode, now promoted from beta to stable.
- Python Code Interpreter tool added to the Tabs.
- HTML/JS Canvas tool added to the Tabs.
- Added attachment icon to the context list if context has attachments.
- Improved audio playback.
- Improved web search.
- Added a thumbnail image to web search results.
- Added a new commands to web search: "extract_images" and "extract_links".
- Added the option "Use raw content (without summarization)" to the web search plugin, which provides a more detailed result to the main model.
- Extended the default maximum result characters to 50,000 in the web search plugin.

2.4.39 (2024-12-09)

- Added "Split Screen" mode (accessible via the switch in the bottom-right corner of the screen), which allows you to work in two windows simultaneously. It is currently experimental (beta). Future updates will include Code Interpreter and Canvas running in tabs.
- Fixed: Language switch.

2.4.38 (2024-12-08)

- Added the ability to select a style for chat display between: Blocks, ChatGPT-like, and ChatGPT-like Wide. New option in the menu: Config -> Theme -> Style...
- Added configuration options for audio input in Settings -> Audio -> Audio Input Device, Channels, and Sampling rate.

2.4.37 (2024-11-30)

- The Query only mode in Uploaded tab has been renamed to RAG.
- New options have been added under Settings -> Files and Attachments:
- Use history in RAG query: When enabled, the content of the entire conversation will be used when preparing a query if the mode is set to RAG or Summary.
- RAG limit: This option is applicable only if 'Use history in RAG query' is enabled. It specifies the limit on how many recent entries in the conversation will be used when generating a query for RAG. A value of 0 indicates no limit.
- Cache: dynamic parts of the system prompt (from plugins) have been moved to the very end of the prompt stack to enable the use of prompt cache mechanisms in OpenAI.

2.4.36 (2024-11-28)

- Added a new command-line argument: --workdir="/path/to/workdir" to explicitly set the current working directory.
- Fix: start image generation in Image mode.

2.4.35 (2024-11-28)

- Docker removed from dependencies in Snap version #82
- Refactored documentation.
- Fix: toggles real-time update hook.
- Fix: missing edit icons.
- Added tokens from attachments to counters if mode == Full context.

2.4.34 (2024-11-26)

- Added a new mode: `Chat with Audio`, with built-in multimodal support for audio input/output. Currently in `beta`, the execution of commands and tools in this mode is temporarily unavailable.
- Added new models: `gpt-4o-audio-preview`, `gpt-4o-2024-11-20`, `chatgpt-4o-latest`.
- Force disabled integration with the native system menu.

2.4.33 (2024-11-26)

- Improved CSS and rendering of file and image lists.
- Added displaying of used attachments in the chat window.

2.4.32 (2024-11-26)

- The "Add URL" option added to the "Attachments" tab allows users to include content from a given website as additional context. Currently, it only supports standard web pages and video transcription for YouTube links. More "web" options will be added in the future.
- Added UTF-8 as default in attachments content text read/write.

2.4.31 (2024-11-25)

- Added an option checkbox `Auto-index on upload` in the `Attachments` tab:

Tip: To use the `Query only` mode, the file must be indexed in the vector database. This occurs automatically at the time of upload if the `Auto-index on upload` option in the `Attachments` tab is enabled. When uploading large files, such indexing might take a while - therefore, if you are using the `Full context` option, which does not use the index, you can disable the `Auto-index` option to speed up the upload of the attachment. In this case, it will only be indexed when the `Query only` option is called for the first time, and until then, attachment will be available in the form of `Full context` and `Summary`.

- Added context menu options in `Uploaded attachments` tab: `Open`, `Open Source directory` and `Open Storage directory`.

2.4.30 (2024-11-25)

- Added instruction to model about mapped data directory in both legacy and IPython code interepreter.
- Updated locales for plugins.

2.4.29 (2024-11-25)

- Added a local IPython interpreter - you can now choose between local and sandbox (Docker) in the plugin settings.
- Added the ability to configure mapped volumes and ports for Docker containers in the plugin settings.
- Optimized and speed-up the Llama-index plugin (inline).
- Checkboxes replaced with Toggle buttons.
- Improved settings dialogs.
- Slight modification of the layout.
- Fix: Dockerfile formatting in Code Interpreter config.
- Fix: experts inline plugin execution.

2.4.28 (2024-11-24)

- Fix: issue #78

2.4.27 (2024-11-24)

- Profile switch fix.

2.4.26 (2024-11-24)

- Fix: issue #77

2.4.25 (2024-11-24)

- Added new plugin: System (OS), with optional sandbox support.
- Execution of system commands moved to the System plugin.
- Improved sandbox/Docker management.
- Enhanced plugin settings.
- Commands renamed to Tools, simplified layout.
- Fix: handling of the Mouse and Keyboard plugin.
- Fix: switching to a new context from a non-chat tab.
- Fix: camera screenshots when the camera is not started.

2.4.24 (2024-11-23)

- Added the ability to use ZIP and TAR archives as attachments (they are now unpacked "on the fly").
- Added the ability to index ZIP and TAR archives (it may be necessary to remove .zip and .tar from the blacklist in the settings).
- Fix: error when uploading to the /data directory using the Upload files button.

2.4.23 (2024-11-23)

- Condensed layout widgets margins.
- Added filename info to provided additional context from attachments.
- Fixed auto-clear in attachments in Assistants mode.

2.4.22 (2024-11-23)

- Added the ability to use attachments with images, audio, and video as additional context. Tip: to enable the use of images as additional context, you need to enable the option: `Files and attachments -> Allow images as additional context`.
- Edit icons in the chat window are enabled by default.

2.4.21 (2024-11-23)

- Added the ability to send additional context from attachments without needing to activate the "Chat with Files" mode. Now, you just need to attach a file, and the additional context from the file will be available in the conversation. More information can be found in the "Files and attachments" section of the documentation.
- Fixed the issue with restarting the agent after it has been forcefully stopped.

2.4.20 (2024-11-22)

- Fixed freeze in Windows installer. #75
- Small fixes.

2.4.19 (2024-11-22)

- Added new LLM providers for Llama-index: HuggingFaceAPI and LocalAI.
- The IPython interpreter now returns local paths to generated files and images (mapped to: %workdir%/ipython/).
- Refactored legacy Agent and Expert modes.
- Fixed Expert mode initialization.
- Added an option in Settings -> Layout to disable animations.

2.4.18 (2024-11-21)

- Refactored kernel status and state handling.
- Fixed: audio input execution.
- Improved output from tools in Agents.
- Added status loading animation indicator.
- Other fixes.

2.4.17 (2024-11-20)

- Refactored kernel and render events.
- Added multiple formatted extra outputs from plugins.
- Added a 'Reset (clear)' option in the context list for clearing the current context.
- Improved agents steps output handling.
- Fixed the audio button.
- Improved the stop event.
- Other small fixes.

2.4.16 (2024-11-19)

- Code Interpreter and Files I/O input/output is now displayed in chat window, using syntax highlight.
- Refactored plugins structure.
- Fix: command execution in Chat in Files mode when no index is selected.
- Fix: missing translations in Agent mode.

2.4.15 (2024-11-18)

- Vision analysis added to all modes.
- Added commands for the Vision plugin: image analysis from an attachment, screenshot analysis, camera image analysis - performed in the background in every mode, without switching to vision mode.
- Improved execution of multiple commands at once.
- Improved integration with IPython and extended instructions for the model.
- Other fixes and improvements.

2.4.14 (2024-11-17)

- Added a 'Loop / evaluate' mode for Llama-index agents, allowing them to work autonomously in a loop and evaluates the correctness of received results using a percentage scale from 0% to 100%, and automatically continue and correct responses below expected value.
- Updated layout: mode and model lists replaced with comboboxes.

2.4.13 (2024-11-16)

- Introduced Code Interpreter (v2) with IPython: Enables session state retention on a kernel and building on previous computations, allowing for a Jupyter-like workflow, which is useful for iterative development and data analysis. Supports magic commands like !pip install for package installation within the session. (currently in beta)

2.4.12 (2024-11-15)

- Added httpx-socks to the dependencies, enabling the use of SOCKS proxies.

2.4.11 (2024-11-15)

- Added a new Llama-index agent: OpenAI Assistant.
- Added proxy settings to the configuration dialog.
- Added more models to the Agent (Llama-index) mode.
- Improved the agents/presets dialog window.
- Disabled the "OpenAI API KEY is not set" Llama-index error when using a local model in Chat with Files mode. You can now use local embeddings, such as Llama3 via Ollama, and use local models without any warnings.

2.4.10 (2024-11-14)

- Added support for Llama Index agents. (beta)
- Introduced a new working mode: Agent (Llama Index).
- Added 3 Llama Index agents: OpenAI Agent, ReAct Agent, and Structured Planner Agent.
- Fixed: passing an embed model to vector stores on indexing.
- Fixed: python3-tk error in snap version.

2.4.9 (2024-11-12)

- Mouse movement has been moved to the PyAutoGUI library for seamless operation.
- Bridge calls have been moved to an asynchronous thread.
- Fixed the DALL-E call in Image mode.
- Speed improvements and optimizations.

2.4.8 (2024-11-12)

- Fix: restoring order of notepads.
- Fix: mouse move offset in Mouse and Keyboard plugin.

2.4.7 (2024-11-11)

- Added a new tool: HTML/JS Canvas, connected to the Code Interpreter plugin. This feature enables real-time rendering of HTML/JS code in a Chromium-based output by using two new Code Interpreter commands: render_html_output and get_html_output.
- Added an SVG icon to the New Tab button.
- Improved the system prompt and execution flow in the Command: Mouse And Keyboard plugin and introduced a new option Auto-focus to the plugin settings (default: False).

2.4.6 (2024-11-10)

- Fixed notepad tab titles in Copy To menu.

2.4.6 (2024-11-09)

- Added a new [+] tab button to the right of the tabs.
- Fixed the issue where the tab name disappeared when creating a new tab.
- Fixed the broken find option (CTRL+F) in the chat window.

2.4.4 (2024-11-09)

- Added a new plugin: Command: Mouse And Keyboard. The plugin allows the model to control the mouse and keyboard, control the cursor position, click mouse buttons, use scroll, type text, and press keys on the keyboard.

2.4.3 (2024-11-08)

- Added button to hide output from tools/plugins.
- Added tooltips on the preset list showing the system prompt of the preset.
- Added restoration of tooltips for chat tabs.
- Removed disruptive tooltips from the plugin settings window.
- Improved reading of current tab information in event reading with speech synthesis.
- Minor improvements.

2.4.2 (2024-11-08)

- Improved commands and tools handling.
- Improved autonomous mode.

2.4.1 (2024-11-08)

- Added titles for conversations in tabs.
- Tab tools added to the tools menu.
- Fix: Reload tabs after changing profile.
- Other fixes.

2.4.0 (2024-11-06)

- Added the ability to have conversations in multiple tabs (right-click on the tab bar to add new tabs via "Add a new chat", similar to a web browser).
- Added the option to add notepad tabs using the right-click context menu and selecting "Add a new notepad".
- Added tab reorganization via drag and drop (you can now arrange tabs freely using the mouse).
- Added the ability to name tabs.

2.3.4 (2024-11-03)

- Support for LaTeX: Mathematical formulas can now be displayed in the chat window using KaTeX library (https://katex.org). This enhancement allows for real-time rendering of complex mathematical expressions.

2.3.3 (2024-11-03)

- Added support for models: OpenAI o1 and o1-mini and SpeakLeash/bielik-11b-v2.2 (Polish LLM: https://bielik.ai).

2.3.2 (2024-08-29)

- Date count fix in SQL query.

2.3.1 (2024-08-29)

- Fixed timezone detection and date conversions in calendar and context search.
- Added icon indicators showing labels in context items in the calendar (to show if any day has a context with the selected label).
- Added custom prompt templates in the input context menu (for saving, editing, and deleting custom prompts).

2.3.0 (2024-08-28)

- Refactored and optimized the chat engine.
- Refactored and optimized command execution.
- Fixed: execution of assistant tool calls.
- Fixed: flow in the expert plugin.
- Fixed: flow in the agent plugin.
- Fixed: console errors in highlight.js.
- Fixed: logging of the context used in the "Chat with Files" plugin.
- Improved the stop command.
- Added a reply stack.
- Other minor fixes.

2.2.33 (2024-08-28)

- Improved model compatibility across different modes.
- Enhanced appending of Llama-index response metadata (sources) to context.
- Added scrolling to the selected item when choosing the current context from the File -> Select current menu.
- Hidden clear preset button.
- Fixed duplicated UUIDs in expert presets.
- Extended context and events debugging.
- Added non-GPT models to Agent and Expert modes.

2.2.32 (2024-08-27)

- Fixed the selection of the chosen index for use in Agent mode if the internal mode for Agent is Llama-index.
- Added the ability to select an index for use by experts.
- Added automatic, in background mode switching if the given mode is not supported by the selected model in Agent mode and for experts (this allows the use of models available only through Llama-index or Langchain in modes such as Agent or Experts).

2.2.31 (2024-08-27)

- Improved behavior of agent and expert calls.
- Improved chat window auto scroll.
- Added a new option "Always continue..." to Agent mode (as opposed to the Auto-stop option).
- Added a new prompt: Config -> Settings -> Prompts -> Agent: continue (always, more steps).

2.2.30 (2024-08-25)

- Added link to Discord server in About -> Discord.

2.2.29 (2024-08-25)

- Fix: Removed 'Index not prepared' monit when index not created yet.

2.2.28 (2024-08-25)

- Fix: Added mechanism to prevent the current index from being deselected when saving settings.
- Added models: llama3.1:70b, llama3.1:405b, and mistral-large.
- Added configuration option: Indexes (llama-index) -> Chat -> Chat mode.
- Split prompt templates context menu into alphabetical submenus.

2.2.27 (2024-08-22)

- Added functionality to hide the date suffix if date period separators are enabled.
- Added a new option: Context -> Show date separators in pinned items on the context list, default: False.

2.2.26 (2024-08-22)

- Added separators for date intervals in the context list.
- Added a new option: Context -> Show context groups on top of the context list, default: False.
- Added a new option: Context -> Show date separators on the context list, default: True.
- Added a new option: Context -> Show date separators in groups on the context list, default: True.

2.2.25 (2024-08-22)

- Improved native API functions handling.
- Improved expert prompts.
- Fixed/improved inline plugin commands handling.

2.2.24 (2024-08-21)

- Added a STOP button that allows stopping indexing at any time.
- Improved conversion of ENUM types when generating a native list of functions for the API.
- Disabled adding functions to the list when the option to execute commands is unchecked.

2.2.23 (2024-08-21)

- Improved context summarization in Llama-index provider.

2.2.22 (2024-08-20)

- Added new API models for "Chat with Files": Google Gemini and Anthropic Claude (see the new section in the documentation: Managing models -> "Using other models (non-GPT)" for more information on how to set API keys)
- Added new local models for "Chat with Files": Llama 3.1, Mistral, Codellama, and Llama2-Uncensored (requires Ollama installed on the local machine - see the new section in the documentation: Managing models -> "Using other models (non-GPT)" for more information on how to enable local model support).
- Added ability for context summarization using Llama-index models (local models can be used to summarize context titles now).
- Added a new option in Config -> Indexes (llama-index) -> Indexing: "Stop indexing on error". If enabled, indexing will stop whenever an error occurs.

2.2.21 (2024-08-20)

- Added inline execution for external function calls.
- Added a stop to file indexing if an error occurs.

2.2.20 (2024-08-20)

- Llama-index upgraded to v0.10.65
- Langchain upgraded to v0.2.14
- OpenAI API upgraded to 1.41.0
- Code syntax highlight changed from Pygments to highlight.js (better performance)
- Added new model: `gpt-4o-mini`
- Added native API function calls by default
- Added functions importer (from plugins) into assistants
- Fixed: input message disappear on command response in stream mode
- Fixed: assistants tool outputs submitting

2.2.19 (2024-05-16)

- Added a newest multimodal model: GPT-4 "omni": https://openai.com/index/hello-gpt-4o

2.2.18 (2024-05-05)

- Fix: prevent crash if no audio to play.

2.2.17 (2024-05-05)

- Fix: Added prevent try to play audio if empty output.
- Disabled playing finish event on audio or voice control enabled.

2.2.16 (2024-05-05)

- Escape key now stops response generation and audio output (if playing).
- Voice control options added to the Audio menu.
- Added cache on disk for generated static audio content.
- Added plugin translations for other languages.

2.2.15 (2024-05-04)

- Added audio output stop on audio input start.
- Added notify about unrecognized command.
- Voice control improvements.

2.2.14 (2024-05-04)

- Added a 'Voice Control (inline)' plugin that allows for voice command control directly during a conversation.
- Added configuration in 'Settings -> Accessibility' for a blacklist of actions available as voice commands.

2.2.13 (2024-05-03)

* Added stretch to dictionary config fields.
* Removed redundant attachments clear event.

2.2.12 (2024-05-03)

- Improved speech recognition.
- Added minimum required length of audio input.
- Added missing translations.
- Fixed settings hooks triggering on profile switch.

2.2.11 (2024-05-03)

- Added a blacklist for events for the voice event description in settings.
- Added a delay to playing audio when describing events.
- Sorted the list of events in the configuration.

2.2.10 (2024-05-03)

- Extended voice control commands list.
- Extended actions and keyboard shortcuts.

2.2.9 (2024-05-02)

- Added more commands to voice control: search for contexts, clear search, add, read and clear calendar memos, context rename.

2.2.8 (2024-05-02)

- Added support for disabled people, including voice control and screen event translation with audio synthesis.
- A new section in Settings called 'Accessibility' has been added with options for assistance: voice control, keyboard shortcut definitions for actions, and screen event translation using audio synthesis.
- A new section called 'Accessibility' has been added to the Documentation.

2.2.7 (2024-05-01)

- A new experimental work mode has been added: 'Experts', which allows the creation of separate background instances with their own instructions that can be consulted for help. See the documentation: 'Work Modes / Experts'.
- Added a new plugin: 'Experts (inline)`.
- Improved the Agent mode by adding the ability to configure and invoke defined Experts.
- Improved the prompts that control the autonomous mode.
- The main prompt controlling the agents has been moved from presets to the application's settings window.

2.2.6 (2024-04-30)

- Added a new model: 'gpt-4-turbo'.
- Vision integrated into Chat mode, without any plugins, if the model supports Vision - currently available for: 'gpt-4-turbo' and 'gpt-4-turbo-2024-04-09'.
- Store importer connected with Logger.
- Fixed: issue with batch unassigning remote files from vector stores.

2.2.5 (2024-04-30)

- Fix: import button handler in Assistants.
- Refactored and improved stores batch actions events handling.
- Added missing translations.

2.2.4 (2024-04-29)

- Batch actions splitted into 'Current store' and 'All stores' in Assistants remote vector stores importer.
- Added sync current/all context menu options to Sync button.
- Improved remote stores handling.

2.2.3 (2024-04-29)

- Fixed Assistants files sync btn.
- Added batch upload files or directory in Assistants vector stores dialog.

2.2.2 (2024-04-29)

- Added 'Application environment' option to Config -> General for setting up custom env vars on app load.
- Added chunk render/parse throttling to prevent high CPU load on huge text chunks.
- Optimized and improved Code interpreter output rendering in stream mode in Assistants.
- Improved run stopping in Assistants.
- Fixed syntax theme and 'Edit' icon restore on profile switch.

2.2.1 (2024-04-28)

- Improved data reload on profile switch.

2.2.0 (2024-04-28)

- Improved code interpreter stream handling in Assistant mode.
- Improved preset plugins reloading on profile switch.
- Fixes and core improvements.
- Updated docs to v2.2.0.

2.1.83 (2024-04-27)

- Index list in 'Chat with files' mode moved to combo-box at bottom.
- Enabled display of remote images in chat output.
- Updated docs.

2.1.82 (2024-04-27)

- Fixed correct store id saving if hidden thread stores on list.

2.1.81 (2024-04-27)

- Improved remote vector stores management.

2.1.80 (2024-04-27)

- Fixed missing selection in Assistant store select combo.
- Fixed store deletion.
- Added tooltip to Stores button.

2.1.79 (2024-04-26)

- OpenAI API upgraded to version 1.23.6.
- Assistants API: added tool: File search and removed deprecated tool: Retrieval.
- Support for remote Assistant Vector Stores added (accessible through the Assistants API).
- Remote Vector Stores management tool introduced (right-click on an Assistant, select 'Edit', then click the database icon to open Remote Vector Stores editor).
- Importer for Remote Assistant Vector Stores and files included.
- Files uploaded in Assistants mode can now be shared between threads using Vector Stores.

2.1.78 (2024-04-25)

- Removed redundant names in block view.
- Scrollbars background integrated with blocks bg.

2.1.77 (2024-04-25)

- Added the "tree" command to the I/O Files plugin to retrieve the full directory tree.

2.1.76 (2024-04-25)

- Chat items now displayed in blocks.

2.1.75 (2024-04-24)

- Copy to clipboard in web view fixed and moved to JS bridge.

2.1.74 (2024-04-24)

- Added syntax highlight choose to Theme menu.
- Syntax highlight style colors integrated with Dark/Light app themes.

2.1.73 (2024-04-24)

- Fixed issue #50 - missing type key in config.

2.1.72 (2024-04-24)

- Fixed context initialization in web view on app load.
- Improved and optimized items rendering.
- Refactored web view CSS.
- Rendering optimization.
- Added real-time items reload and append.
- Added theme CSS to scrollbars.
- Added JS console debug to Logger window.

2.1.71 (2024-04-23)

- Added 160+ prompt templates from "Awesome ChatGPT Prompts" repository - to use them, right-click on the input or system prompt textarea and select the "Paste from template..." option.

2.1.70 (2024-04-22)

- Added web view CSS reload on theme switch.
- Fixed time display CSS in web view.

2.1.69 (2024-04-22)

- Added html buffer to speed-up context initialization in web view.
- Fixed items append in web view in agent mode.

2.1.68 (2024-04-22)

- Fix: CPU usage on multiple items.

2.1.67 (2024-04-22)

- Optimized items rendering.

2.1.66 (2024-04-22)

- Improved content formatting in HTML export.

2.1.65 (2024-04-22)

- Improved save as in Web View output.
- Added save as text and save as html separated options.

2.1.64 (2024-04-22)

- Improved real-time CSS reloading.

2.1.63 (2024-04-22)

- Improved rendering output.
- Updated CSS.

2.1.62 (2024-04-22)

- Added default lexer for code syntax highlighting if no language provided.

2.1.61 (2024-04-21)

- Added real-time HTML formatting of text chunks in stream mode.

2.1.60 (2024-04-21)

- Fixed "Save as" option in WebEngine view.
- Added hiding of empty groups on search.
- Updated formatting and CSS.

2.1.59 (2024-04-20)

- Added output renderer: WebEngine/Chromium – with full HTML/CSS/JS support. (Experimental)
- Added a new option to choose a render engine in Settings -> General -> Rendering engine to select between WebEngine / Chromium mode and the Legacy mode (old, with simple markdown CSS support only - for compatibility reasons and troubleshooting).
- Added syntax highlighting to code blocks – issue #43.
- Added a new option in Settings -> Layout -> Code syntax highlight to select a highlight theme.
- Added a new option in Settings -> Context -> Convert lists to paragraphs to enable/disable converting ul and li into p for better copying.
- Copy to clipboard and regenerate response icons are now visible by default.

2.1.58 (2024-04-19)

- Fixed Markdown rendering if the content starts with a code block.
- Added a "Switch to created profile" option in the profile creation/duplication dialogs.
- Updated the locale.
- Added a donate option.

2.1.57 (2024-04-17)

- Fixed new ctx create on empty profile if previously group was selected.

2.1.56 (2024-04-17)

- Added a new menu option: Clear history + groups.
- Improved profile switching.

2.1.55 (2024-04-17)

- Improved result handling from Indexer tool.

2.1.54 (2024-04-17)

- Added data loaders reload and loaders config reload in real time on profile switch.

2.1.53 (2024-04-17)

- Added a new tool: Indexer, for manually indexing files, contexts, and web content, as well as managing indexes. (beta)

2.1.52 (2024-04-15)

- Added real-time output from code interpreter in Assistants mode.
- Added new option in Settings > Context: Show code interpreter output.

2.1.51 (2024-04-14)

- Added RMB context menu to the New... context button.
- Added filesize info to attachments and uploaded files lists.
- Improved Assistants API stream handling.

2.1.50 (2024-04-14)

- Fixed selection update after group rename.
- Added img resolution and filesize info in image preview dialog title.
- Added filesize info in text editor dialog title.

2.1.49 (2024-04-14)

- Improved context group handling.

2.1.48 (2024-04-14)

- OpenAI API upgraded to version 1.17.1.
- Added a stream mode to the Assistant mode (beta).
- Fixed an issue where context group ID were not copied when duplicating context.

2.1.47 (2024-04-12)

- Added auto-creating new context in current group if the group is active.
- Added sending input also with Ctrl+Enter in Shift+Enter mode.
- Extended light style markdown CSS.

2.1.46 (2024-04-12)

- Improved code blocks formatting.
- Improved attachment copy/paste.

2.1.45 (2024-04-11)

- Added the ability to paste images as attachments using Ctrl+V in chat input and to copy-paste images in the draw/painter.
- Added a "Copy to Clipboard" option in code blocks.
- Introduced a new model: gpt-4-turbo-2024-04-09.

2.1.44 (2024-04-11)

- Optimized "Find" option.
- Added copy database and copy data dir checkbox options in profile duplicate dialog.
- Added ability to reset current active profile.
- Added missing translations.

2.1.43 (2024-04-11)

- The "Find" option has been extended and added to all text fields.
- Fixed menu reload issue when switching profiles.
- Other small fixes and improvements.

2.1.42 (2024-04-10)

- Improved profile switching.
- Added interpreter internal files to excluded when indexing.

2.1.41 (2024-04-10)

- Added support for multiple profiles (separated config, settings and workdir), menu: Config -> Profile...
- Added context groups for context items organizing, menu: File -> New context group...
- Added ability to change the model on response edit and regenerate.
- Added real-time updating number of notepads.
- UI/theme fixes.

2.1.40-post1 (2024-04-08)

- Llama-index core downgraded to 0.10.13 due to issue #41

2.1.40 (2024-04-08)

- Added "Find" option - text search and highlighting of the found phrase in the chat and notepad window, available via CTRL + F or right-click menu / Find... (beta)
- Added GitHub's Monaspace font family: Argon, Krypton, Neon, Radon, Xenon (reset CSS to defaults or manual replacement in CSS files is required for older versions of application to take effect).
- Added an "Restore Default CSS" option in the "Config / Edit CSS" menu.
- Added the ability to use custom .ttf and .otf fonts (just place them in %workdir%/fonts directory).

2.1.39 (2024-03-26)

- Added "Change working directory..." option in "Config" menu.
- Improved handling reply from multiple commands at once.
- Tools have been moved to separate parts of the app.
- Added the ability to create custom Tools, with an example tool in the "examples" directory.

2.1.38 (2024-03-20)

- Added new Tools menu, with: Media player, Transcribe audio/video files, Image viewer, Text editor and Code interpreter.
- Improved tools.

2.1.37 (2024-03-19)

- Added generation of audio transcriptions from audio/video files.

2.1.36 (2024-03-19)

- Added split screen view (output/edit) to code interpreter.

2.1.35 (2024-03-19)

- Added text file editor and image viewer.
- Improved video player and code interpreter.

2.1.34 (2024-03-18)

- Added video player.
- Fixed audio input message append in agent mode.

2.1.33 (2024-03-18)

- Code interpreter enabled even if interpreter plugin is disabled.
- Added "Connect to the Python code interpreter window" config option in Code interpreter plugin.
- Added Monospace font support - issue #37.
- Improved raw JSON command outputs.

2.1.32 (2024-03-17)

- Global prompts config moved to a separated settings section "Prompts".
- Added auto-clear option in code interpreter.
- Added "Use extra context output" config option in "Settings -> Context".

2.1.31 (2024-03-17)

- Improved code interpreter integration with sandbox/docker.
- Added display of hidden files in Files explorer.
- Added "Use as image" context menu action in Files explorer.

2.1.30 (2024-03-16)

- Improved real-time Python code interpreter: added edit mode and whole code execution (history).

2.1.29 (2024-03-16)

- Added real-time Python code interpreter (<> icon), connected with the Code Interpreter plugin's input and output.
- Improved plugin command execution.

2.1.28 (2024-03-15)

- Fixed local commands handling in Assistant API.
- Fixed system prompt replace after mode changed on app start.
- Improved related links rendering in chat output.
- Added RPM limit config option for embeddings API.
- UI improvements.

2.1.27 (2024-03-13)

- Added the "find" command to the Files I/O plugin for searching for files and directories.
- Added "Append metadata" config option in Chat with files plugin.
- Added missing locales to translations.

2.1.26 (2024-03-13)

- Added async command execution in History and Chat with files plugins.
- Improved plugin prompts.
- Improved command execution in autonomous mode.
- Improved logging and debugging.

2.1.25 (2024-03-13)

- Llama index logging improvement.

2.1.24 (2024-03-13)

- Added AzureOpenAI embeddings provider.
- Added metadata append to indexed context from DB.

2.1.23 (2024-03-12)

- Added configuration options for embeddings model and provider in Settings -> Llama-Index.
- Added plain-text responses in web_open commands.
- Improved web_open commands in Web Search plugin.
- Added a check to determine if the provided URL is an image when sending URLs within text.

2.1.22 (2024-03-12)

- Added custom metadata appending to indexing documents for web/external loaders.

2.1.21 (2024-03-12)

- Refactored commands core and command handling to a JSON schema-based config.
- Fixed calls in the API plugin.
- Fixed web_urls in the Web search plugin.
- Fixed adding new notes in the Calendar plugin.

2.1.19 (2024-03-11)

- Added "Force exclude files" config option to force exclude extensions in: Settings -> Llama-index -> Indexing (enables file extension exclusion even if a data loader for the extension is available).
- Added custom metadata appending to indexing documents via list in: Settings -> Llama-index -> Indexing.
- Added base API endpoint config arguments in models params.

2.1.18 (2024-03-10)

- Added horizontal splitters and min width for lists in settings dialogs - issue #30.
- Added audio icon display config option in Settings -> Context.

2.1.17 (2024-03-10)

- Improved audio handling.
- Added options to context debug.
- Added context meta load on demand if not on list.

2.1.16 (2024-03-10)

- Added "merge" icon in context responses (allows merging selected response with the previous response).
- Added display of Llama-index used sources in the context response.
- Fixed word-wrap in code blocks - issue #27.
- Fixed context visibility when filters are applied and new ctx is created in the background.
- Improved audio playback status indication for responses.

2.1.15 (2024-03-09)

- Added Edit mode in chat output with options: read response (audio), copy to clipboard, regenerate response, edit and regenerate, and delete context item.
- Added dynamic max output tokens calculation in completion mode - issue #28.
- Added config option "Modes to auto-index" in Llama-index settings.
- Fixed typo in gpt-4-turbo-preview model config args for Llama-index provider - issue #29.

2.1.14 (2024-03-09)

- Added restoration of "All counters" checkbox status in calendar on tab init.
- Added new line with Shift+Enter in input if sending mode is "Enter" - issue #26.

2.1.13 (2024-03-08)

- Fixed context auto-indexing in autonomous/agent mode - issue #23.
- Added AzureOpenAI provider for Llama.
- Added configuration option for the model in query file/web options in plugins.
- Added removal of temporary index from memory after use.

2.1.12 (2024-03-08)

- Added config option for max requests per minute (RPM) in Settings -> Model - issue #23.
- Added config for continue prompt in Settings -> Agent.
- Added config option for custom API endpoint URL in Settings -> General - issue #21.
- Added Allow context item delete option in Settings -> Context.
- Added checkbox All counters in filters in Calendar view - when disabled, calendar displays filtered record counters instead of all records.
- Improved the reset to default option in Settings to persist all non-settings config options.

2.1.11 (2024-03-07)

- Fixed search query if multiple labels selected
- Added prevent from saving empty notes in calendar
- Added display empty dates in DB Viewer if timestamp = 0

2.1.10 (2024-03-06)

- Added "DB Viewer" in Developer menu.
- Added label color filter in context list - issue #23
- Added context item delete option - issue #23
- Added presets for plugins - issue #23
- Fixed and improved autonomous agents run - issue #23
- Chinese language added to locales - issue #24

2.1.9 (2024-03-04)

- A new option has been added to the "Web Search" plugin: "web_index_query", which allows for indexing a web and external content in a temporary index (in memory) and quickly querying its content. Works similar to "query_file" command.
- The "read_file" command has been expanded to handle multiple files at once.
- Added "Log and debug events" option to Developer settings.

2.1.8 (2024-03-03)

- A new option has been added to the "Files I/O" plugin: "query_file", which allows for indexing a file in a temporary index (in memory) and quickly querying its content. This enables the use of Llama-index for fast querying or adding context from individual files. From now on, you can ask questions about individual files without indexing them to the main vector store. See the "Querying single files" in "Modes -> Chat with files" section in a README for more info.
- The "read_file" command has been expanded to return just the exact file content, without wrapping it in JSON.
- Improved prompt for preparing queries in the Chat with files plugin.
- Updated docs.

2.1.7 (2024-03-03)

- Enhanced the Llama-index plugin with additional options for automatically refining queries before requesting additional context from Llama-index.

2.1.6 (2024-03-03)

- Added Twitter/X data loader.
- Added check if message is not a reply when calling Llama-index using a plugin.
- Added list of Agent indexes auto-reload after Llama-index list has changed.
- Disabled anonymized telemetry in Chroma by default.

2.1.5 (2024-03-02)

- Recompiled SVG resources.
- Fixed file explorer update after file commands execute.
- Updated docs.

2.1.4 (2024-03-02)

- PySide6 downgraded to 6.4.2 due to a bug in signals emitting.

2.1.3 (2024-03-01)

- Added data loaders: Bitbucket, ChatGPT Retrieval Plugin, SQL Databases, GitHub, Microsoft OneDrive.
- Added indexing configuration tab in Web Search plugin.

2.1.2 (2024-03-01)

- Added data loaders: Google Drive, Google Docs, Google Gmail, Google Calendar, Google Sheets, Google Keep.
- Added data loaders: Video/Audio and Images, with local model and API modes.
- Added Whisper (local hosted) model to Audio Input providers, available only in Python version.

2.1.1 (2024-02-29)

- Llama-index settings split into tabs.
- Added keyword argument settings for data loaders.
- Added data loaders for RSS, XML sitemaps, IPYNB notebooks, and XML and HTML files.
- Added audio notes - simply use speech recognition in the Notepad tab to append transcribed audio to the Notepad.
- Added an icon for quick toggling of video camera capture.
- Added a Video menu.
- Moved video capture toggle checkboxes to the Video menu.

2.1.0 (2024-02-28)

- Upgraded Llama-index from the legacy version 0.9 to the current version 0.10 (note: upgrading PyPi version of PyGPT to 2.1.0 requires reinstallation in a virtual environment to remove old library versions).
- Upgraded Langchain to 0.1.9.
- Upgraded PyInstaller (and compiled versions) to version 6.
- Removed LlamaHub and restored support for Python 3.12.
- Fixed imports for Llama-index in the compiled versions.
- Added icons for quickly toggling audio input/output.
- UI fixes.

2.0.171 (2024-02-27)

- Added core for loading web resources (not only webpages) to Llama-index.
- Added a data loader for transcriptions of YouTube videos (simply ask for index a YouTube video URL using the "Command: Web Search" plugin - the "web_index" command must be enabled).
- Fixed the "Initial token count exceeds token limit" error in Chat with files mode.
- Added missing translations for dictionary items in settings dialogs.

2.0.170 (2024-02-26)
- Added "simplified" mode to the "Audio Input" plugin, now set as default, allowing for faster and simpler speech recognition without requiring any configuration. Meanwhile, all current options have been moved to "advanced" mode.
- Improved microphone support in Snap version.
- Improved file browser update after performing file operations.
- Added check for file availability for indexing in the file browser.

2.0.169 (2024-02-25)

- Improved and refactored debugging

2.0.168 (2024-02-25)

- Added auto index file on read option in the "Files I/O" plugin
- Improved debugger
- Fixed Bing speech recognition

2.0.167 (2024-02-25)

- Fixed empty API key dialog.

2.0.166 (2024-02-25)

- Enabled the use of data loaders from Llama-index in the "Files I/O" plugin for reading files, which allows for reading not only text files using the plugin's "read_file" command.
- Added a command for indexing files to the "Files I/O" plugin.
- The "self_loop" plugin has been internally renamed to the "agent" plugin.
- Fixed the display of pinned contexts when the indexed filter is enabled.

2.0.165 (2024-02-25)

- Added search provider: Bing.
- Added audio output providers: Google and Eleven Labs.
- Added audio input providers: Google, Google Cloud, and Bing.
- Added a new plugin: "Command: API calls" which allows defining commands for API calls to external services.
- Added an "Excluded file extensions" configuration option in Settings -> Llama-index.
- Fixed audio input listener initialization in the Audio Input plugin.

2.0.164 (2024-02-24)

- Refactored the Audio Input, Audio Output, and Google Web Search plugins to support multiple audio and search engine providers.
- Introduced the option to add your own audio input, output, and web search providers.

2.0.163 (2024-02-23)

- Added storage of indexed urls
- Added info about used library versions in About dialog

2.0.162 (2024-02-23)

- Index data storage migrated from JSON to DB.
- Added add/remove to index context menu on ctx list.
- Added last indexed info on ctx list.
- Added option "Replace old document versions in the index during re-indexing" in "Settings -> Llama-index".
- Refactored, optimized and improved file and context indexing.
- Fixed context search SQL query.

2.0.161 (2024-02-22)

- Added the option to search within conversation content, not just by titles (disabled by default, can be enabled in: "Settings -> Context -> Search also in conversation content, not only in titles").
- Added a "download" subdirectory for files downloaded in Assistants mode and an option to configure the directory in: "Settings -> Files and attachments -> Directory for file downloads")
- Added logging of options available in: "Config -> Debug" to the logger window, not just to the console.
- Added descriptions in the models editor.

2.0.160 (2024-02-21)

- Added auto-replacement of sandbox download links in Assistants with links to downloaded local files.
- Added setup for Poetry.
- Added connection timeout config option to Google Search plugin.
- Fixed recursion error on app exit.
- Updated dependencies.

2.0.159 (2024-02-21)

- Added workdir placeholders to images stored from Assistants.
- Added tooltips on modes list.
- Added force synchronous command call if more than one command to execute.

2.0.158 (2024-02-21)

- Improved Assistants: added annotations handling, added async function outputs submit
- Fixed creating function duplicates on re-import existing Assistant

2.0.157 (2024-02-20)

- Added webpage (URL) indexing using Llama-Index; new options were added to the Google Search plugin: the command "web_index" for indexing a specified page (just ask to index any URL) and "Auto-index", which enables automatic indexing of URLs used by the plugin (disabled by default).
- Decreased the default chunk size in the Google Search plugin.
- Assistant run creation and message sending moved to an async thread.
- Improved files receive from Assistants.

2.0.156 (2024-02-19)

- Added disable SSL verify option for search engine calls (not only for search results) in "Google Search plugin" - issue #20.
- Updated certifi to 2024.2.2.
- Added custom launcher, custom plugin, custom vector store, custom LLM provider and custom data loader code examples in "examples" directory in repository.
- Updated documentation.

2.0.155 (2024-02-18)

- Importing assistants and assistant files (from API) has been moved to an asynchronous thread.
- Upgraded OpenAI API to 1.12.0.

2.0.154 (2024-02-18)

- Fixed hidden user input in Assistants mode
- Fixed system prompt change when switching to Assistants mode without selected Assistant

2.0.153 (2024-02-17)

- Added image display in the chat window when a file received from the Assistant is an image
- Added color labels in color select combo boxes
- Added dialog window position save and restore (+ option in Settings to enable/disable this feature)
- Added command "web_url_raw" to the Google Search Plugin which allows you to get the raw HTML/txt content from web pages (without summary)
- Fixed CSS backup path
- Updated documentation

2.0.152 (2024-02-16)

- Improved updater, added direct links to new releases
- Improved command execute system prompt
- Added "cwd" command to "Files I/O" plugin
- Added automatic current working directory passing to "sys_exec" command prompt in "Code Interpreter".
- Added "Refresh" option in File Explorer
- Added monit about camera connect in Snap version with command to execute
- Disabled MenuRoles on MacOS-affected menu items - issue #19

2.0.151 (2024-02-16)

- Improved integration between Chat and Vision when functions are provided

2.0.150 (2024-02-16)

- Added a new command for the "Google Web Search" plugin - "get_urls", which allows retrieving a full list of searched URLs and passing it to the model for further use.
- Implemented execution of "functions" defined in the API and their full integration with modes: "Chat" and "Assistant". For more details, see the new section in the documentation: "Functions and commands".
- Fixed the tray icon change while the model is generating a response.
- Improved stopping of agents using "Stop".

2.0.149 (2024-02-15)

- Fixed and improved Assistants API (file handling, run handling, token usage, stopping)
- Added vertical canvas resolutions in Painter
- Added brush settings restore in Painter
- Added log file viewer in Debug menu
- Added log level switch in runtime (Developer settings)
- Added functions defining and handling in Presets (works in ChatCompletion API)
- Improved presets and assistants editor dialogs
- Improved debugger and Developer settings

2.0.148 (2024-02-07)

- Fixed input/output timestamps in renderer
- Added aspect ratio check and fit on image open in Painter tool

2.0.147 (2024-02-04)

- Added @ tags for quick access to context from previous discussions in the Context History plugin (must be enabled in plugin settings, default: disabled).
- Added automatic passing of the day of the week in the Context History and Real Time plugins.

2.0.146 (2024-02-03)

- Disabled online loaders for Llama-index if compiled version is detected (they work properly only in Python version)

2.0.145 (2024-02-03)

- Enabled the use of the Context history plugin in inline mode (without the need for the Command execution option).
- Added automatic provision of the current date in the Context History plugin if the Real-Time plugin is disabled.
- Added the ability to count contexts within a specified date range.
- Added an option to limit the maximum number of contexts retrieved per query.
- Enabled the "undo" option in the painter when clearing and opening a new image.

2.0.144 (2024-02-02)

- Improved Command: Context history plugin prompts

2.0.143 (2024-02-02)

- Added a new plugin: Command: Context history (calendar)
- Added a new feature to context list: Pin on top
- Added Minimize to tray on exit option

2.0.142 (2024-02-01)

- Added tray notification on scheduled task run.
- Added tray notification on Agent goal achieved.
- Small fixes.

2.0.141 (2024-02-01)

- Improved window hide/restore from tray icon
- Removed capture disable/enable options and Start minimized from settings dialog

2.0.140 (2024-02-01)

- Fixed preset duplicating

2.0.139 (2024-02-01)

- Improved the generation of names for preset duplicates.
- Added background updates check (with an option in settings to disable it).
- Added dialog to accept license terms at the first launch.
- Fixed calendar content save on month change.
- Small UI improvements.

2.0.138 (2024-02-01)

- Added "Duplicate" option to current * (default) presets
- Added "Restore to default" option to current * (default) presets
- Added "Minimize to tray" settings option
- Fixed preset selection when clicked on list with RMB
- Improved calendar records count on month break

2.0.137 (2024-01-31)

- Improved Agent / Autonomous mode handling
- Added "Remove file from index" option in file-explorer
- Other small fixes

2.0.136 (2024-01-31)

- Improved integration of Agent mode with llama-index sub-mode
- Updated documentation and locales

2.0.135 (2024-01-31)

- Fixed issue where the agent would stop after the first run in Agent mode when the number of iterations was set to non-infinity.
- Added "Agent" section in the settings, with options to select the sub-mode to use (Chat, Completion, Langchain, Llama-index, etc.) and an option to choose the index to use.
- Implemented a feature to display the current run status in Agent mode.

2.0.134 (2024-01-30)

- Updated models list

2.0.133 (2024-01-30)

- Added the ability for autonomous image generation using the "DALL-E (inline)" plugin in "Agent mode".
- Added a "Redo" option to the "Painter/Draw" tool.
- Enabled Vision in "Agent mode".

2.0.132 (2024-01-30)

- Experimental: Added new working mode: "Agent (autonomous)", which works similarly to the "Autonomous Mode" plugin but as a separate mode with editable system prompt presets.
- Renamed "Autonomous Mode" plugin to "Autonomous Mode (inline)".
- Fixed real-time system prompt tokens calculation update.
- Fixed preset saving in editor.
- Updated icons' fill color.

2.0.131 (2024-01-30)

- Fix: set a limited height for list items in editable lists within the options.
- Added new plugin: `System Prompt Extra`: plugin appends additional system prompts (extra data) from a list to every current system prompt. You can enhance every system prompt with extra instructions that will be automatically appended to the system prompt.
- Added a list of prompts for define and use in the `Autonomous Mode` plugin (the previous division into "normal" and "extended" prompts has been replaced with an unlimited, user-editable list).
- Added textarea stretch on dialog resize in dictionary options editor
- Introduced a new `duplicate` option in the context list (it copies the selected context to a new one).

2.0.130 (2024-01-29)

- Added status indicator in Tray - issue #12
- Added icons to tabs, tray and context menus
- Added "Files and attachments" section in Settings

2.0.129 (2024-01-29)

- Added Material Design SVG icons (by Google)
- Added icon resources compiler script
- Improved file explorer

2.0.128 (2024-01-29)

- Added improvements to RMB context menu handling in Files
- Improved generation of names for duplicated files

2.0.127 (2024-01-28)

- Added ability to auto-store (in the "%workdir%/upload" directory) local copies of attachments uploaded from outside the workdir.
- Added an option in the general settings dialog: "Store attachments in the workdir upload directory", default: True
- Added an option for storing everything in the "data" directory: "Store images, capture and uploads in data directory", default: False

2.0.126 (2024-01-28)

- New features in file explorer (Files tab):

- Create empty file
- Create directory
- Upload files to selected directory
- Download file or directory
- Duplicate file or directory
- Use as attachment
- Ask for read this file
- Copy working/system path to input and clipboard

2.0.125 (2024-01-28)

- Fixed chars encoding in save text to file option

2.0.124 (2024-01-28)

- Fixed Vision checkbox visibility when switching to plain-text mode
- Added RMB context menu option "Save as..." to selected text for saving selected text (or whole document) to a text file
- Added remembering the last used directory for uploading attachments/saving files

2.0.123 (2024-01-27)

- Standarized models names
- Added "Recursive directory indexing" option in "Settings / Llama-index" (default: False)

2.0.122 (2024-01-27)

- Added new OpenAI model released on 2024-01-25: "gpt-4-0125-preview"
- Added ability to add custom offline data loaders for Llama-index
- Improved debugging and logging, added debug levels, and introduced a new section "Debugging and Logging" in the documentation
- Moved inline styles to external CSS classes
- Renamed "pygpt_net.provider.llm" to "pygpt_net.provider.llms"

2.0.121 (2024-01-27)

- Fixed query error when using Raw mode in Chat in files mode with enabled plugins
- Added bridge / abstraction layer between modes
- Hidden assistant checkbox in preset editor

2.0.120 (2024-01-26)

- Fixed font-size restore when switching between plain text and markdown renderer
- Added RMB context menu option "Clear files" to input Attachments tab.

2.0.119 (2024-01-25)

- Fixed list selection loose after async command execution
- Added "Start minimized" option in settings
- Updated docs

2.0.118 (2024-01-25)

- Added "Show tray icon" option in "Settings"
- Added launcher arguments
- Updated docs
- Small fixes

2.0.117 (2024-01-24)

- Added a new feature to the "Command: Files I/O" plugin - "Get and upload file as attachment". It lets automatically read files from "data" folder and send them as attachments - just ask to send a file as an attachment and the model will send it to itself (like sending an image for analysis). Option is disabled by default in the plugin's settings.
- Added a new plugin: "Command: Serial port / USB". This plugin lets you read and send any commands to USB ports (so you can control robots now ;) ).
- Made it possible to use URLs for images when not using the vision feature.
- Fixed saving and reloading prompts in Assistant mode.
- Removed display of empty external images from the output.

2.0.116 (2024-01-24)

- Fixed font color in light themes
- Fixed layout display on Windows systems
- Theme, Language and Debug menus moved to Config menu - issue #10
- Added Developer section in Settings
- Fixed list options editing
- Refactored code and updated tests

2.0.115 (2024-01-22)

- Added token recalculation after the attachments list is changed and upon copy/pasting into the input field.
- Added descriptions to the settings.
- Improved the "Copy selected text to..." option.

2.0.114 (2024-01-21)

- Fixed broken CSS on Windows
- Added support for Vector Store databases: "Chroma", "Elasticsearch", "Pinecone" and "Redis" (beta)
- Added config options for selecting and configuring Vector Store providers
- Added ability to extend PyGPT with custom Vector Store providers
- Added commands to the "Vision (inline)" plugin: get camera capture and make screenshot. Options must be enabled in the plugin settings. When enabled, they allow the model to capture images from the camera and make screenshots itself.
- Added "Query index only (without chat)" option to "Chat with files" mode.
- Added stream mode support to query index mode in "Chat with files".

2.0.113 (2024-01-20)

- Added %workdir% placeholder to attachments and images files paths storage for more flexibility in moving data between environments
- Refactored base plugin options handling

2.0.112 (2024-01-19)

- Fixed image variants slider in image mode
- Fixed vision checkbox visibility
- Fixed user directory path when handling different directories/symlinks
- Added config option for disable opening image dialog after image generate
- Added Scheduled tasks entry in taskbar dropdown
- Added * (asterisk) indicator after date modified if newer than last indexed time
- Date of modified in file explorer changed to format YYYY-MM-DD HH:MM:SS

2.0.111 (2024-01-19)

- Fixed: opening files and directories in Snap and Windows versions
- Fixed: camera capture handling between mode switch
- Added: info about snap connect camera in Snap version (if not connected)
- Added: missing app/tray icon in compiled versions

2.0.110 (2024-01-19)

- Fixed bug: history file clear on ctx remove - issue #9
- Vision inline allowed in modes: Langchain and Chat with files (llama-index)
- Event names moved to Event class

2.0.109 (2024-01-18)

- Fixed bug: float inputs value update behaviour - issue #8
- Added: plugin description tooltips - issue #7
- Added: focus window on "New context..." in tray - issue #13
- Added: Ask with screenshot option to tray menu - issue #11
- Added: Open Notepad option to tray menu - issue #14

2.0.108 (2024-01-16)

- Added confirmation dialogs on indexing
- Added missing translations
- Updated dependencies: openai, cryptography, certifi, urllib, jinja

2.0.107 (2024-01-16)

- Fixed model change in vision plugin
- Fixed completion API error
- Removed deprecated davinci and replaced with 3.5-turbo-instruct model

2.0.106 (2024-01-15)

- Added "undo" action in drawing window
- Refactored locale handling
- Improved locale and theme overriding

2.0.105 (2024-01-15)

- Extended "Theme" menu with new options
- "Index (llama-index)" mode changed name to "Chat with files" mode
- Refactored theme controller
- Fixed model config providing to llama-index mode

2.0.104 (2024-01-14)

- Langchain support refactored and fixed
- Added editor for models
- Added options for customize arguments and environment vars passed to langchain and llama providers
- Added context length limit in modes: langchain and llama-index
- Models configuration integrated between GPT/Langchain/Llama-index

2.0.103 (2024-01-13)

- Llama-index mode, now integrated with context from the database, streaming, plugins, and more, can be utilized as any other mode and is no longer limited to just querying indexes.

2.0.102 (2024-01-13)

- Added presets for llama-index mode
- Added custom system prompt in llama-index mode
- Added option to use online data loaders in settings
- Added hiding of chat footer in non-chat tabs
- Added missing libraryfor epub data loader

2.0.101 (2024-01-13)

- User interface improvements
- Added help tips

2.0.100 (2024-01-13)

- Llama-index integrated with database - you can now indexing all conversations and use them as additional context
- Added support for multiple indexes at once
- Added configuration for llama-indexes
- Data-dir in home directory changed name from 'output' to 'data'

2.0.99 (2024-01-12)

- Added configuration sections in settings dialog
- Added translations to file explorer
- Langchain migrated to langchain community
- Config patchers moved to separated classes
- Fixed some UI issues

2.0.98 (2024-01-12)

- UI fixes

2.0.97 (2024-01-11)

- DALL-E config fix

2.0.96 (2024-01-11)

- Added integration with Llama-index (experimental): you can now index/embed files from the data directory and use them as additional data in context.
- Added an inline plugin for Llama-index (allows providing additional context from Llama-index in any chat mode).
- Built-in file loaders: text files, pdf, csv, md, docx, json, epub, xlsx
- Added a new mode: Index (llama-index), it is a mode for querying the index directly.
- Improved DALL-E 3 configuration, 16:9 images set as default, added config option for standard and HD mode.
- Improved drawing mode.

2.0.95 (2024-01-10)

- Added capture from camera feature in image drawing mode

2.0.94 (2024-01-10)

- Improved drawing: added canvas, image load/save, store and restore drawing on start and more
- Improved dict option editor: added type fields

2.0.93 (2024-01-10)

- Added painter (simple image drawing) - it allows to quick draw images and provide them into vision model
- Small UI fixes

2.0.92 (2024-01-10)

- Fixed bug with system prompt append from main window
- Added editor for list options (via RMB -> Edit...)
- User/AI names and temperature removed from main window
- UI fixes

2.0.91 (2024-01-09)

- Added new plugin: Crontab / Task scheduler: the plugin provides cron-based job scheduling - you can now schedule tasks/prompts to be sent at any time using cron-based syntax for task setup.
- Added combo-box to configuration option types
- Fixed incorrect contexts count in date calculation when timezone is different from UTC
- Fixed incorrect paragraph formatting after command DIV in the last line of response
- Other small fixes and improvements

2.0.90 (2024-01-08)

- Fixed config load defaults bug
- Refactored configurations core
- Small UI fixes

2.0.89 (2024-01-07)

- Added secondary, extended prompt to autonomous mode configuration (it allows quick switching between standard and more extended reasoning).
- Fixed user input disappearance in history when appending inside autonomous mode.

2.0.88 (2024-01-07)

- Added color labels to context items (you can now mark item on list with 'Set label color...' context menu option)
- Improved storage of notepad and calendar items

2.0.87 (2024-01-07)

- Fixed paragraph formatting and font color.
- Implemented automatic database backup creation prior to all migrations.
- Added functionality to search contexts using both fields in a single query: search string and date range.

2.0.86 (2024-01-07)

- Fixed window state saving on exit
- Added focus on input on creating new context

2.0.85 (2024-01-07)

- Improved Autonomous mode
- Added synchronous command execution from plugins, which now allows the Autonomous mode to use the output of other plugins (such as access to the websearch engine, access to filesystem and code execution or generating images in DALL-E) and retrieve and use response from them internally
- Fixed system prompt for Autonomous mode: resolved the problem of command generation interruptions between responses, and incorporated the capability for user to send additional instructions and update goals in real-time
- Fixed loss of selection on the modes and models list in Autonomous mode

2.0.84 (2024-01-06)

- The "Self-Loop" plugin has been completely redesigned, improved, and renamed to the Autonomous Mode, which enables the launch of autonomous reasoning in a looped AI-with-AI conversation mode.
- A system prompt option has been added to the plugin configuration with instructions for the autonomous mode.
- Auto-stop after goal is reached option is added also to plugin settings.

2.0.83 (2024-01-06)

- Added calendar, day notes (memos) and color labels
- Added context search filter by date (days selected in calendar)
- Added 'clear' icon in search input
- Small fixes and improvements

2.0.82 (2024-01-05)

- Plain text append scroll fix

2.0.81 (2024-01-05)

- Added Edit context menu option in models list (JSON file edit)
- Fixed link color in light theme css
- Fixed font color restore when switching to plain text

2.0.80 (2024-01-05)

- Added Plain Text checkbox at the bottom of chat window
- Removed datetime append when copying text to notepad
- Improved plain-text mode

2.0.79 (2024-01-05)

- Fixed bug: notepad save/load content from DB

2.0.78 (2024-01-05)

- Improved markdown formatting
- Output text renderers moved to separated modules
- Added plain-text render option in settings
- Vision inline disabled in unsupported modes (Image generation and Assistant) and hidden in Vision (always enabled here)
- Updated OpenAI Chat Completion API

2.0.77 (2024-01-05)

- Fixed inline plugins disable when pausing command plugins by unchecking 'Execute commands' checkbox
- Added real-time image appending into chat window when generating images in Image Generation mode
- Added automatic creation of context after deletion from the list when no other context is selected
- Updated docs

2.0.76 (2024-01-04)

- Markdown post-process changed from markdown extension to BS4 html parser.

2.0.75 (2024-01-04)

- Improved default theme for markdown styling
- Added "Check for updates on start-up" config option in updater dialog
- Added custom CSS stylesheets editor in Config menu
- Added password show/hide in password/secret input fields (click on the right corner of input reveals plain text)
- Fixed JSON files editing (added data reloading after save to prevent overwrite on exit)
- Added tests for plugins and extended tests for core/controllers

2.0.74 (2024-01-03)

- Fixed timestamp position when appending input to the chat window.
- Extended the Markdown parser with an extension for converting "li" to "p" (it allows copying lists bullets with Ctrl-C).
- Added notebook titles to the "Copy to..." context menu.
- Added an "Open file" option in the file explorer and attachments list.

2.0.73 (2024-01-03)

- Fixed vision prompt appending
- Improved inline vision: removed the "keep" option and added a vision indicator with a checkbox for quickly enabling/disabling the vision model if needed (will auto-enable when an image is provided if the plugin is active)

2.0.72 (2024-01-02)

- Fixed markdown / HTML formatting
- Added config option "Use color theme in chat window" in main settings
- Added configuration for system prompt to use with vision in Vision Inline plugin settings

2.0.71 (2024-01-01)

- Added Markdown and HTML formatting in the chat window (beta version)
- Added the display of images, files and URLs internally in the chat window
- Added fully integration with images, attachments, and files inside the chat window
- Added vision capture and image generation directly into the chat window (via the DALL-E Inline plugin)
- Improved audio speech generation (eliminated the silence that prematurely truncated the previous audio output before generating the next speech synthesis)

2.0.70 (2023-12-31)

- Added commands part strip from speech generation
- Added img check for links sending to vision in auto mode

2.0.69 (2023-12-31)

- Added a new plugin: GPT-4 Vision (inline - in any chat). Plugin integrates vision capabilities with any chat mode, not just Vision mode. When the plugin is enabled, the model temporarily switches to vision in the background when an image attachment or vision capture is provided.

2.0.68 (2023-12-31)

- Added a new plugin: DALL-E 3 Image Generation (inline) which enables image generation in any chat and mode, seamlessly in the background - just request an image from any model, like GPT-4, and it will generate it "inline."
- User Interface optimizations.
- Added type hints to the code.
- Fixed issue with sending attachments in Assistant mode.

2.0.67 (2023-12-30)

- Notepad restore fix

2.0.66 (2023-12-30)

- Added "Rename" option to Notepad Tabs (via RMB)
- Improved language switching

2.0.65 (2023-12-30)

- Image generation, vision capture and assistants run listeners moved to async threadpool
- Added saving state of scroll value in chat window
- Added search string save and restore
- Added input text save and restore
- Updated tests and code cleanup
- Added translations (only main app): DE, ES, FR, IT, UA
- Small fixes and optimizations

2.0.64 (2023-12-29)

- Commands and plugins execution moved to asynchronous native QT threadpool
- Improved theme switching
- Improved debugger window
- Small fixes and optimizations

2.0.63 (2023-12-28)

- Fixed context append bug
- Improved layout resize

2.0.62 (2023-12-28)

- Fixed date display on context list
- Added mode display in context tooltips
- Added "Contexts list records limit" config option in settings
- Added tokens calculator tooltips

2.0.61 (2023-12-28)

- DALL-E multiple messages in one context fix

2.0.60 (2023-12-28)

- Message append fix

2.0.59 (2023-12-28)

- Storage of context has been changed from JSON files to SQLite database, automatic migration is included - you will not lose your existing data after update :)
- Added "Search" functionality for contexts and a search input at the bottom.
- Updated contexts now automatically land at the top of the list.
- Added the ability to store more information about each context.
- Integration of searching, utilizing SQLite, and vector database search coming soon.
- Improved DALL-E image generation.

2.0.58 (2023-12-27)

- Code refactor
- Added stream response mode to Vision mode
- Added token calculation to Langchain mode
- Optimizations and small fixes

2.0.57 (2023-12-26)

- Added threadpool for async workers handling
- Fixed segmentation fault error on app exit
- Refactored class names

2.0.56 (2023-12-25)

- Reorganized project structure
- Relative imports changed to absolute imports
- Updated core paths
- Fixed plugin response hanging
- Fixed recursion problem in error logger
- Fixed platform module name

2.0.55 (2023-12-25)

- Fixed max model ctx tokens

2.0.54 (2023-12-25)

- Updater new version check moved to post-setup
- Fixed default assistant preset
- Added link to Snap Store to updater window

2.0.52 (2023-12-25)

- Added DPI option in settings: enable/disable and scale factor
- Fixed font-size change with mouse scroll on notepad window

2.0.51 (2023-12-25)

- Added auto-content truncate before tokens limit exceeded
- Added dynamic max tokens switching
- Fixed real-time tokens calculation

2.0.50 (2023-12-25)

- Fixed async logging
- Updated tests

2.0.49 (2023-12-24)

- Fixed crash on async command call
- Fixed UI issues
- Added data providers
- Refactored classes

2.0.48 (2023-12-23)

- Added custom errors handler with logging.

2.0.47 (2023-12-23)

- Added "Copy to Notepad/Input" to the context menu for selected text.
- Added an option to read selected text using speech synthesis in the Notepad window.
- Added a configurable number of notepads.
- Refactored UI classes.
- Added a thread-safe plugin debugger/logger.
- Fixed UI issues.

2.0.46 (2023-12-22)

- Improved tokens calculation (added extra tokens from plugins to real-time calculation)
- Added new context menu option to selected text in output/input: "Read with speech synthesis"

2.0.45 (2023-12-22)

- Added "light" themes
- Fixed tokens calculation
- Updated tests

2.0.44 (2023-12-21)

- Fixed language switch in plugins settings

2.0.43 (2023-12-21)

- PyGPT publicated at Snap Store: https://snapcraft.io/pygpt
- Added link to Snap Store in app menu

2.0.42 (2023-12-20)

- Audio output library changed to `PyGame` mixer (instead of `PyDub`)
- Added `PyGame` as a dependency and removed `PyDub` and `simpleaudio` dependencies
- Added audio input/output stop immediately on plugin disable

2.0.41 (2023-12-20)

- Added "sandbox" feature to the Code Interpreter – it allows the use of any Docker image as an environment for code and commands execution.
- Context auto-summary moved to an async thread.
- Command execution moved to an async thread.

2.0.40 (2023-12-19)

- Totally improved material themes support
- Added QSS-configurable themes to chat output window and code highlighter
- Added ability to overriding styles per any theme in user directory
- Fixed UI layout elements
- Fixed DPI issues in Windows systems

2.0.39 (2023-12-18)

- Fixed dialogs closing with the Esc key
- Added an indicator for "can append to context in this mode"
- Added functionality to fetch filenames from the API when importing files uploaded to Assistants
- Enabled switching to newly created Assistants after creation
- Optimized class structure

2.0.38 (2023-12-18)

- Multi-language support added to plugins
- Implemented feature to override locale and CSS files by overwriting them in the home directory
- Bug fixes

2.0.37 (2023-12-18)

- Fixed prompt typing in real-time on the right toolbox
- Added toolbox font change option
- Simplified custom plugins and wrappers for LLMs application (see Docs for more details)
- Bug fixes
- UI fixes

2.0.36 (2023-12-18)

- Improved plugins handling
- Added Event Dispatcher and event-based system

2.0.35 (2023-12-18)

- Fixed lists focus lost and selection disappearing
- Added clickable links to API keys pages
- Added scroll position store in notepads
- Auto auto-switch to preset after create
- Assistants files and thread moved to external classes
- UI fixes
- Code fixes

2.0.34 (2023-12-17)

- Added option "Lock incompatible modes" to prevent mixing of incompatible contexts.
- Added PyPI link to the About menu.

2.0.33 (2023-12-17)

- Fixed dialog's save/update button handlers
- Fixed uploaded files reload in assistants
- Fixed focus loss on assistants list
- Fixed Output Files header
- Fixed UI
- Added delete confirmation to dictionary options
- Added context change lock before response generation
- Added allowed types for contexts
- Added a link to the documentation in the menu
- Added saving state of opened advanced options in plugins
- Disabled mouse scroll on sliders

2.0.32 (2023-12-16)

- Added real-time font-size change with CTRL + mouse scroll in input, output and notepad windows
- Increased allowed font-size to 42
- Fixed line display in plugin settings

2.0.31 (2023-12-16)

- Speech recognition (Whisper) small fixes, optimization and improvements
- Added advanced internal options to speech recognition config

2.0.30 (2023-12-15)

- Speech recognition and synthesis fixes and improvements
- Fixed and improved speech recognition via Whisper
- Fixed and improved voice synthesis via OpenAI TTS and MS Azure
- Added new options to speech recognition: magic words, stop words, prefix words, auto-send and wait for response
- Added new more intuitive voice input/output control panel in UI

2.0.29 (2023-12-14)

- Added config and plugin options getters/setters
- Changed logo
- Small UI fixes

2.0.28 (2023-12-14)

- Added new hidden Credentials/API Key field type (asterisks against plain-text)
- Simplified presets editing
- Fixed Assistants function management
- Improved UI
- Improved settings editing

2.0.27 (2023-12-14)

- Added specified "url open" command to Web Search plugin
- Added additional advanced options to above plugins
- Improved Web Search and Commands Execution
- Improved updater and config patcher
- Improved command execution logging

2.0.26 (2023-12-13)

- Added advanced config options for plugins
- Added additional file operations to Files I/O plugin

2.0.25 (2023-12-13)

- Added advanced settings
- Added clear on capture config option
- Added capture quality config option
- Added launcher shortcuts
- Improved plugins
- Improved WebSearch
- Optimized commands response
- Fixed UI issues

2.0.24 (2023-12-12)

- Fixed empty string in tokens calculator
- Added attachments reset before auto-capture from camera

2.0.23 (2023-12-12)

- Improved python code execution

2.0.22 (2023-12-12)

- Fix: env API KEY name for Langchain mode

2.0.21 (2023-12-12)

- Simplified assistant configuration
- Added assistant configuration validation
- Improved UI
- Improved language switcher

2.0.20 (2023-12-11)

- Improved Assistants API
- Added assistant uploaded files storage
- Added assistant uploaded files management
- Added assistant remote functions management
- Fixed "open in directory" option on Windows in DALL-E image generation
- Improved attachments and file upload management
- Improved UI and more

2.0.19 (2023-12-10)

- Optimized DALL-E prompt generator helper

2.0.18 (2023-12-10)

- Config fix

2.0.17 (2023-12-10)

- Small fixes

2.0.16 (2023-12-10)

- Added multiple cameras config
- Added DALL-E prompt generation RAW MODE ON/OFF switch
- Improved camera handling

2.0.15 (2023-12-10)

- Added camera release / disable on camera off

2.0.14 (2023-12-10)

- Added real-time video capture from camera in "Vision" mode

2.0.13 (2023-12-10)

- Fixed path resolving in "open in directory" option on Windows OS
- Added real-time apply of "layout density" (after "save changes" in Settings)
- Default "layout density" changed to 0
- Updated locale

2.0.12 (2023-12-09)

- Improved system commands execution

2.0.11 (2023-12-09)

- Small fixes

2.0.10 (2023-12-09)

- Updated locale

2.0.9 (2023-12-09)

- Added Command: Custom Commands feature; plugin allows to easily create and execute custom commands
- Added new features to Command: Files I/O: downloading files, copying files and dirs, moving files and dirs

2.0.8 (2023-12-08)

- Improved Web Search plugin

2.0.7 (2023-12-08)

- Improved code execution with Code Interpreter / Files I/O plugins

2.0.6 (2023-12-08)

- Added layout density configuration

2.0.5 (2023-12-08)

- Added support for external CSS
- Added custom fonts support
- Improved material theme support

2.0.4 (2023-12-08)

- Added configuration options for plugins: Files I/O, Code Interpreter
- UI fixes

2.0.3 (2023-12-07)

- Python code execution fix

2.0.2 (2023-12-07)

- Added python command template settings
- Added layout state restore
- Refactored settings
- Improved settings window
- Bugfixes

2.0.1 (2023-12-07)

- Fixed settings dialog initialization
- Fixed models.json migration
- Added enter key behaviour settings
- Added font size settings for input and context list
- Added ctx auto-summary settings
- Added python command plugin settings

2.0.0 (2023-12-05)

New features in version 2.0.0:

- Added support for new models: GPT-4 Turbo, GPT-4 Vision, and DALL-E 3
- Integrated Langchain with support for any model it provides
- Assistants API and simple assistant configuration setup
- Vision and image analysis capabilities through GPT-4 Vision
- Image generation with DALL-E 3
- File and attachment support including upload, download, and management
- New built-in notepad feature
- Multiple assistants support
- Command execution support
- Filesystem access allows GPT to read and write files
- Asynchronous (stream) mode added
- Local Python code interpreter that enables code execution by GPT
- System command executions directly from GPT
- Voice synthesis provided via Microsoft Azure TTS and OpenAI TTS (Text-To-Speech)
- Voice recognition provided by OpenAI Whisper
- Automatic summarization of context titles
- Upgraded Web Browser plugin
- More precise token calculation functionality
- Added output markup highlight
- Improved UX
- Bug fixes
- Plus additional enhancements and expanded capabilities

0.9.6 (2023-04-16)
- Added real-time logger
- Improved debug mode

0.9.5 (2023-04-16)
- Added web plugin (adds access to the Internet using Google Search and Wikipedia)
- Added voice output plugin (adds voice synthesis using Microsoft Azure)

0.9.4 (2023-04-15)
- Added plugins support

0.9.2 (2023-04-12)
- Added theme color settings
- Small UI fixes

0.9.1 (2023-04-11)
- Added organization key configuration
- Added config version patching

0.9.0 (2023-04-09)
- Initial public release