ThinkInAIXYZ/deepchat

AI Summary: Integrate the MCP Router Marketplace API into the DeepChat application, allowing users to easily discover, install, and manage MCP servers directly within the app's GUI. This involves adding a new Marketplace section with filtering capabilities, handling API requests for server manifests, and automatically configuring the selected servers within DeepChat's settings.

AI Summary: Add support for the Poe API as a new model provider in the DeepChat application. This involves integrating Poe's OpenAI-compatible API, requiring configuration of the API base URL and authentication method (Bearer Token using POE_API_KEY), and ensuring compatibility with the existing application architecture.

AI Summary: Add support for the GLM-4.5 language model to the DeepChat application's BigModel.cn provider. This involves adding a new entry for `glm-4.5` to the list of available models within the provider's configuration, allowing users to select and utilize this model like others.

AI Summary: Enhance the DeepChat application to support previewing of various code formats (SVG, Mermaid, etc.) within a separate window during chat sessions, similar to the existing Artifact preview functionality. This involves potentially modifying the markdown renderer library used by the application.

AI Summary: Implement MCP Progress functionality in the DeepChat application. This involves integrating the MCP Progress specification to provide real-time progress updates during model computations and tool calls, enhancing user experience and providing better feedback on long-running tasks.

AI Summary: Add support for the Voyage AI API to the DeepChat application. This involves integrating their OpenAI-compatible API, handling authentication with a Bearer token, and potentially adding specific UI elements for configuring Voyage AI models. The primary focus is on embedding models, but chat models should also be considered.

AI Summary: Add support for the ModelScope API as a new provider in the DeepChat application. This involves integrating the ModelScope API (which uses an OpenAI-compatible endpoint and Bearer token authentication) into the existing multi-model architecture, allowing users to select and utilize ModelScope's models within the DeepChat interface.

AI Summary: Add support for Hyperbolic.ai as a new model provider in the DeepChat application. This involves integrating their OpenAI-compatible API, requiring configuration of the API base URL and authentication method (Bearer Token with API key). The implementation should be consistent with the existing model provider integrations.

AI Summary: Add support for Tokenflux.ai as a new Large Language Model (LLM) provider in the DeepChat application. This involves integrating their API, likely OpenAI-compatible, requiring API key authentication and configuration within the DeepChat settings. The task includes verifying the API endpoint and authentication method from Tokenflux.ai's official documentation.

AI Summary: The task is to add support for the Cephalon Cloud API as a new model provider in the DeepChat application. This involves integrating the Cephalon Cloud API, which is OpenAI-compatible, into the existing multi-model architecture, handling authentication (Bearer Token), and ensuring seamless user experience within the DeepChat interface.

AI Summary: Add support for the BurnCloud API to the DeepChat application. This involves adding a new provider configuration option, handling authentication (likely Bearer Token), and potentially adjusting existing code to accommodate the BurnCloud API's specific endpoint structure and response format. The exact API details need to be confirmed with BurnCloud's documentation.

AI Summary: Add support for the Groq LLM provider to the DeepChat application. This involves implementing a new provider module that utilizes Groq's OpenAI-compatible API, handling authentication with a Bearer token, and integrating it into the existing user interface.

AI Summary: Implement MCP sampling support in the DeepChat application. This involves adding client-side capability declaration, a user interface for reviewing and approving/rejecting sampling requests, handling `sampling/createMessage` requests and responses, and supporting various message content types (text, image, audio). The implementation should adhere to the MCP specification and allow users to configure model preferences and storage of approved models or auto-approval strategies.

AI Summary: Implement a "human-in-the-loop" feature for the DeepChat MCP (Model Context Protocol) flow. This involves allowing users to intervene and take action when issues like authentication failures or insufficient permissions occur during MCP operations. The design should allow for seamless continuation of the process after user intervention, similar to the Linear.app example provided. This requires designing the user interface for intervention, handling various error scenarios, and integrating the intervention mechanism into the existing MCP flow.

AI Summary: Debug a bug in the DeepChat application where pasting an image causes a "illegal base64 data at input byte 4" error when using the Gemini 27b model. The task involves analyzing the provided error message and screenshot to identify the root cause of the image pasting failure within the DeepChat application's image handling and Gemini 27b model interaction.

AI Summary: Implement a feature in the DeepChat application to add API request rate limiting to prevent exceeding API provider limits and resulting errors. This involves adding a configurable setting to control the rate of requests and implementing logic to pause requests when the limit is reached.

AI Summary: Refactor the DeepChat application's settings views to improve performance and user experience by consolidating all settings into a single, unified page accessible from the shell, instead of being spread across multiple tabs.