Deepchat, from ThinkInAIXYZ, is an open-source, local-first AI agent desktop client. It supports the Model Context Protocol (MCP) and integrates with a variety of AI models and services, including cloud APIs like OpenAI, Gemini, and Anthropic, alongside locally deployed Ollama models. Built with TypeScript, Electron, and Vue.js, Deepchat provides a desktop application runtime for Windows, macOS, and Linux. It functions as an MCP client, designed to enable semantic workflows and automation by connecting to other tools.

What does Deepchat connect to?

Deepchat connects to various AI models and services, functioning as an AI agent desktop client and an MCP client. It supports cloud APIs like OpenAI, Gemini, and Anthropic, alongside local Ollama models. The client is compatible with AI services using OpenAI, Gemini, and Anthropic API formats. Confirmed compatible services include DeepSeek, SiliconFlow, Grok, Claude, Hugging Face, Kimi, LiteLLM, MiniMax, Mistral, OpenRouter, OpenWebUI, Perplexity, Qwen, Requesty, StabilityAI, Together, and X.

Core features and tools

Deepchat offers "Skills" as callable tools, installable from folders, ZIP files, or URLs. These Skills can contain task instructions, reference files, assets, and optional scripts. The client also includes "inMemory services" for essential functions like code execution, web information retrieval, and file operations. Built-in Skills cover diverse workflows such as generative art, code review, document collaboration (DOCX, PDF, PPTX, XLSX), frontend design, git commit messages, infographic syntax, and MCP building. These capabilities allow for versatile AI agent interactions.

Who is it for?

Deepchat targets individuals and enterprise teams, as well as developers and businesses, who require an AI agent desktop client or chat platform. It serves those looking to integrate AI capabilities into their daily tasks, development processes, or content creation workflows. The client's local-first design and support for various AI models make it suitable for users prioritizing data control and flexibility. This is not the right fit if the primary need is an MCP server.

How to get started

The authentication model primarily involves providing API keys for connecting to AI models, which are configured within the application settings. For MCP setup, adding JSON settings for specific MCP services is required within Deepchat's settings.

Use cases

• Acting as a daily assistant for answering questions, providing suggestions, and assisting with writing and creation.
• Serving as a development aid for code generation, debugging, and technical problem-solving.
• Facilitating content creation through copywriting, creative inspiration, and content optimization.

Deepchat provides a versatile AI agent desktop client for connecting to many AI models and services. Its local-first approach and support for the Model Context Protocol make it a flexible tool for various applications. The client’s "Skills" and "inMemory services" extend its utility across different workflows, from development assistance to content creation.