0.41.0

Highlights

🗣️ Voice Control & Speech-to-Text: Talk to AiderDesk!

Unleash a new level of productivity with Voice Control and Speech-to-Text support! AiderDesk can now hear your voice. If you have Gemini or OpenAI providers configured, you can enable this feature to create your prompts directly using your voice. We've started with these providers due to their excellent real-time models, ensuring a smooth and responsive user experience, and are exploring options for other providers.

✅ Multiselect for Task Sidebar: Manage Tasks with Ease

Managing numerous tasks just got a whole lot simpler! The task sidebar now supports multiselect. Use CTRL + Click (or long press) to activate multiselect mode, allowing you to archive or delete multiple tasks at once. This is a huge time-saver for keeping your project history organized.

🎨 Introducing the "Fresh" Color Theme

Give your AiderDesk a vibrant new look! We've added a brand new "Fresh" color theme, designed by the Gemini 3 Pro model itself. If you appreciate light themes, we believe you'll find this new addition both refreshing and visually appealing.

Other Changes

• Added resize functionality to the todo window for improved usability.
• Enhanced internal handling for Aider run tool, correctly auto-approving questions and denying execution of bash commands when Agents uses the tool.
• Corrected a parameter name for an internal helper tool, improving robustness when LLMs use invalid tool arguments.
• Ensured missing parameters are correctly added to providers when the "Disable streaming" option is used.
• The read_file tool message now includes line and limit information for better context.

💡 We’d love to hear your thoughts! Your feedback helps us improve AiderDesk - feel free to share any suggestions or issues. Happy coding! 🚀

0.40.0

Highlights

🌡️ Enhanced Model Configuration: Temperature & Max Tokens at Your Fingertips

Gain unprecedented control over your AI's behavior! You can now configure temperature directly for each model in the Model Library, allowing for precise fine-tuning of response creativity. We've even added default optimal values for specific models like Gemini. Plus, the Max Tokens setting in agent profiles is now optional, defaulting to the model's configured value. This streamlines your workflow, as you no longer need to adjust these settings every time you switch models.

🚀 Deeper Aider Integration with Model Library Settings

We've significantly improved how Aider Desk's Model Library integrates with the Aider process. Aider now correctly utilizes the costs and temperature settings directly from your Model Library, reducing the need for separate Aider-specific model configuration files. Our goal is to make the Model Library your single source of truth for all model parameters. If there are other parameters you'd like to see configurable here, please let us know!

✨ Streamlined New Task Creation

Creating new tasks is now more intuitive and efficient! When you initiate a new task, it immediately appears as [Untitled] in your task sidebar, allowing you to easily save it by simply setting a name. This also brings a significant advantage for Git Worktree users, as the new branch will be created based on the task name, providing a clearer and more organized project structure.

🤝 New Provider Support: MiniMax

We're excited to announce direct support for the MiniMax provider! MiniMax offers a "Coding Plan" subscription, making it a compelling option for those seeking cost-efficient yet capable model. Our implementation utilizes Anthropic's compatible API, ensuring a smooth integration. Definitely worth checking out for your coding tasks!

Other Changes

• Added maxResults parameter to the grep tool, with a default of 50, for more controlled search results.
• The "Show/Hide all files" state now persists across tasks within the same project.
• Improved utilization of abort signals and enhanced interruption handling for a more robust experience.

💡 We’d love to hear your thoughts! Your feedback helps us improve AiderDesk - feel free to share any suggestions or issues. Happy coding! 🚀

0.39.1

Hotfix

• fixed issue with project files not showing in the Context Files

Complete release notes for version 0.39.0 can be found here.

0.39.0

Highlights

📖 Enhanced file_read Power Tool: Smarter Context Management

The file_read power tool has been significantly enhanced to give your Agent more precise control over file content. It can now read only specific parts of a file and display line numbers, optimizing the context window by providing the AI with only the most relevant information and improving its capabilities when analyzing code.

🛡️ Fine-Grained Control for bash Power Tool: Auto-Approval & Denial Patterns

Take full control over your Agent's shell commands! The bash power tool now includes new settings to define patterns for commands that are automatically allowed or always denied. This allows you to build a trusted execution environment, giving you greater confidence in the commands the Agent runs, even when in auto-approve mode.

✅ Task-Level Auto-Approval for Flexible Workflows

We've moved the auto-approve setting from the Agent profile to the individual task level. This means each task can now have its independent auto-approval configuration. Some tasks can be fully automated for hands-off execution, while others can remain under close watch, offering maximum flexibility for diverse workflows.

🗄️ Archive & Search Tasks: Organize Your Project History

Keep your task list tidy without losing valuable work! You can now archive tasks that you want to keep for future reference but don't need in your active list. Access them anytime, and use the new search functionality in the task sidebar to easily find any task you're looking for, making project history management a breeze.

⚙️ Disable Streaming for Improved Provider Compatibility

Some local or specialized providers don't work optimally with streaming responses. We've added a new option to disable streaming per provider, and even per specific model. This ensures smoother and more reliable interactions with a wider range of AI services.

⚡ New Provider Support: GPUStack & Enhanced OpenAI Compatibility

AiderDesk continues to expand its ecosystem:

• GPUStack Provider: We've added support for GPUStack, allowing you to leverage powerful local GPUs for running models and seamlessly connecting to AiderDesk.
• OpenAI Compatible Reasoning: For those using OpenAI-compatible providers, we've added reasoning effort support, enabling them to utilize this advanced capability for better AI performance.

🚀 Performance Boosts for a Snappier Experience

Expect a faster, more responsive AiderDesk! We've implemented several performance improvements:

• Optimized Message Loading: Faster loading of large numbers of messages.
• Enhanced Autocompletion: Optimized autocompletion, making the prompt field feel quicker and more fluid.
• Faster All Files Loading: Loading of all files is now more efficient.
  These ongoing optimizations contribute to a smoother and snappier overall working experience.

Other Changes

• Improved agentic flow for providers that stop right after a tool call, ensuring smoother operations.
• Fixed occasional white screen errors when using the compacting feature.
• Provider overrides in settings now accurately reflect the current state of provider settings.
• Custom commands now work properly in Agent mode when used within the browser.
• Improved auto-scrolling behavior while responses are being generated.
• Added missing provider parameters when generating commit messages.
• A new task branch is now created automatically when using Git worktree functionality.
• Added a copy button to the subagent prompt for convenience.
• Dropping files now clears them immediately instead of waiting for an event from Aider.
• Fixed provider options handling when multiple OpenAI Compatible providers are configured.

💡 We’d love to hear your thoughts! Your feedback helps us improve AiderDesk - feel free to share any suggestions or issues. Happy coding! 🚀

0.38.1

Hotfix

• fixed occasional error when using /compact command
• fixed issue with response getting stuck in Aider modes in some cases

Complete release notes for version 0.38.0 can be found here.

0.38.0

Highlights

🌳 Git Worktree Integration for Isolated Task Development

Revolutionize your workflow with Git Worktree integration for tasks! Each task can now operate within its own isolated Git worktree, allowing you to:

• Work on Changes in Isolation: Develop features or fix bugs without affecting your main branch.
• Effortless Merging & Squashing: Easily merge your changes back into the main branch or squash them for a cleaner history.
• Quick Reverts: Unhappy with the results? Revert changes with a single action to try a different approach.
  This feature also makes it incredibly easy to experiment with the same prompt across different agents or models. To further enhance this, we've added the ability to save prompts for future task executions and duplicate tasks to explore multiple variants of your work. Stay tuned for even more workflow enhancements related to tasks!

✨ Enhanced Task Workflow & Management

We've introduced several quality-of-life improvements for task management:

• Save & Duplicate Prompts: Prepare a prompt, save it for later, and execute it when ready. Duplicate tasks to explore different approaches or models.
• Export Tasks: Easily share or archive your work by exporting tasks to Markdown or as an Image.
• Copy Task ID: Quickly reference tasks with a new action to copy their unique ID.
• Welcome Message: A friendly welcome message now greets you in empty task views.
• Performance & Stability: Improved performance related to message rendering and fixed issues with dropping all files after settings changes.

🔄 Model Library Refresh

Keeping your model list up-to-date is now simpler! We've added a Refresh action to the Model Library, allowing you to reload available models with ease.

☁️ Azure Provider Enhancements

The Azure provider has received significant attention:

• Fixed Reasoning Models: Reasoning models are now fully functional with Azure.
• Reasoning Effort Support: Added support for setting the reasoning effort, giving you more control over AI behavior.

Other Changes

• Improved performance related to rendered messages for a smoother experience.
• Empty "Thinking" blocks are now omitted from messages for cleaner output.
• The browser client now loads the current question and triggers an event when answering.
• Avoid creating a new task when one is already in a new (unsaved) state.
• Fixed weak model prefix not being set correctly when different from the main model.
• Remember the prompt field value when switching between tasks.
• Avoid selecting a task when only opening its context menu.

💡 We’d love to hear your thoughts! Your feedback helps us improve AiderDesk - feel free to share any suggestions or issues. Happy coding! 🚀

0.37.1

Hotfix

• fixed issue with no responses when using Requesty provider
• todo tools are now properly used with new task management

Complete release notes for version 0.37.0 can be found here.

0.37.0

Highlights

✨ Introducing Task Management: Revolutionize Your Workflow!

We're thrilled to introduce a brand-new Task Management system, replacing the previous session functionality with a more integrated and powerful approach. Now, each of your projects can effortlessly keep track of an unlimited number of tasks (formerly sessions). This allows you to:

• Organize All Your Work: Easily archive previous work or pick up exactly where you left off on any task.
• Boost Productivity: Run multiple tasks simultaneously! Let the Agent tackle one task in the background while you plan or work on another. This dramatically increases your efficiency and helps you get more done.

This is just the beginning for Task Management; we have ambitious plans to further enhance this feature, including individual model and mode settings per task. Stay tuned for more exciting developments!

🧠 Auto-Compacting Conversation History

Say goodbye to overflowing context windows! AiderDesk now supports auto-compacting of your conversation history when token usage exceeds a configurable threshold. This intelligent feature automatically summarizes your long conversations during Agent runs, ensuring the AI always has the most relevant context without being overwhelmed. You can set your preferred token threshold or disable it if you prefer manual context management. Thanks to @pcfreak30 for the suggestion.

Other Changes

• Improved API key handling from environment variables in Agent mode. Thanks to @ohnum for getting this done.
• Fixed an issue where the weak model prefix was not being set correctly when different from the main model.
• Added the possibility to lock the weak model, providing more control over model selection. Thanks to @MotherSoraka for the suggestion.

💡 We’d love to hear your thoughts! Your feedback helps us improve AiderDesk - feel free to share any suggestions or issues. Happy coding! 🚀

0.36.0

Highlights

✨ Enhanced Model Library & Usage Dashboard

We've made it easier to manage and find your AI models!

• Hidden Models by Default: In the Model Library, less frequently used "Hidden" models are now hidden by default, decluttering your view. A new checkbox allows you to easily show or hide them as needed.
• Sortable Tables: Both the Model Library and the Usage Dashboard tables now feature sorting capabilities. This means you can quickly organize models and usage data by criteria like token count, cost, or other relevant metrics, making it simpler to find the information you need. (thanks @Slach for the suggestion)

📁 Organized Custom Commands with Subfolders

Your custom commands just got a major organizational upgrade! You can now arrange your custom commands within subfolders inside your .aider-desk/commands/ directory. For example, place a command in review/last-commit.md and invoke it directly in your prompt using review:last-commit. This brings a more structured and intuitive command experience, reminiscent of popular coding tools.

⬆️ Drag & Drop Agent Profile Sorting

Personalize your Agent workflow even further! You can now easily reorder your Agent profiles in the settings using intuitive drag-and-drop functionality. Move your most frequently used or favorite agents to the top of the list for quicker access.

🚀 Seamless Auto-Approval for Subagents

For those who prefer an uninterrupted workflow, the main agent's auto-approve setting is now automatically passed down to subagents. This means when "auto-approve" is enabled, even if a subagent is configured to use tools that normally require confirmation, they will execute without interruption, ensuring a truly "You Only Live Once" (YOLO) coding experience. (thanks @pcfreak30 for the suggestion)

Other Changes

• The "About" tab in Settings now correctly opens when clicking the "Update is available" icon.
• Fixed an issue with an incorrect tool definition for the AiderDesk MCP server.

⚠️ Important: Behind-the-Scenes Refactoring for Project Tasks

We've undertaken significant refactoring behind the scenes, moving closer to a "Tasks within a project" architecture. As part of this, your session data located in .aider-desk/sessions has been migrated to .aider-desk/tasks. While these changes have been thoroughly tested, there might be specific workflow-related bugs we didn't manage to catch. If you encounter anything not working as expected, please help us improve by creating an issue.

💡 We’d love to hear your thoughts! Your feedback helps us improve AiderDesk - feel free to share any suggestions or issues. Happy coding! 🚀

0.35.0

Highlights

🚀 AI SDK v5 Migration & Enhanced Tooling Capabilities

We've upgraded our underlying AI SDK from v4 to v5, laying a robust foundation for future support and exciting new integrations with various AI providers. This migration immediately unlocks powerful new capabilities:

• Native Web Search for OpenAI Models: You can now configure your OpenAI models to utilize their native web search functionality, enabling them to fetch real-time information directly.
• Unrestricted Gemini Search Grounding: Gemini models can now leverage search grounding without limiting the use of other tools, providing more flexible and powerful multi-tool interactions.

🚫 Clearer Tool Denied Feedback

To provide immediate and intuitive feedback, we've updated the icon displayed when a tool is denied. This visual cue makes it much clearer why a tool did not execute, helping you understand and adjust your workflow more efficiently.

Other Changes

• Correctly using OPENAI_ environment variables for the Z.AI Coding Plan provider in Aider modes.
• Removed _API_KEY environment setting for Aider modes when it's not explicitly defined in the Provider.
• The /model command now correctly opens the agent model selector when in Agent mode.
• Fixed an issue with duplicated parts of the message appearing during streaming.
• Properly using overridden custom token cost values for models.
• Now showing an informational message when "Export to Image" of messages is not available due to virtualized rendering.
• Properly reconstructing thinking messages after loading a session, even when there are multiple message parts.

💡 We’d love to hear your thoughts! Your feedback helps us improve AiderDesk - feel free to share any suggestions or issues. Happy coding! 🚀