- Atomic Chat
Atomic Chat
What is Atomic Chat?
Atomic Chat is a free, open-source desktop and mobile app for running local LLMs and connecting MCP servers directly to your models. It acts as an MCP host: add a server once and its tools are available in every chat — local or cloud.
How to use MCP with Atomic Chat
-
Download from https://atomic.chat (macOS, Windows, Linux, iOS, Android)
-
Open Settings → MCP Servers → Add
-
Paste any MCP server config (name, transport: STDIO / HTTP / SSE, command or URL)
-
Start chatting — the model picks and calls tools automatically
Works with any server from mcp.so and the official MCP registry.
Key features
- MCP host - connect multiple MCP servers; in app log viewer for tool calls
- Local models - Llama, Qwen, DeepSeek, Gemma, Mistral and 1000+ more from Hugging Face
- Cloud providers - OpenAI, Anthropic, Mistral, Groq, MiniMax (bring your own key)
- OpenAI-compatible API -
http://localhost:1337/v1for agents, CLIs, and IDE plugins - Privacy - with local models, prompts and tool calls stay on your machine
- Built for agents - integrates with Kilo Code, OpenCode, Goose, and others.
Use cases
- Connect Notion, GitHub, Google Drive or web search to a local model
- Run agents and coding tools on your own hardware via the local API
- Mix local inference with cloud models without reconfiguring MCP servers
FAQ
Which MCP transports are supported?
STDIO, HTTP, and SSE.
Does my data go through Anthropic or OpenAI?
With a local model, tool calls go from your machine to the MCP server you connected — not through an AI provider.
Is it free?
Yes. Open-source under Apache 2.0.
Where to download?
https://atomic.chat
Recommend Clients
View AllThe AI Code Editor