Repolode
Submit
ThinkInAIXYZ

ThinkInAIXYZ/deepchat

An open-source, local-first AI agent desktop client supporting MCP, Skills, ACP agents and remote control from messaging apps.

Open-source alternative to:ChatGPTClaude Desktop

View on GitHubWebsite
Stars
6,372
Forks
759
Open issues
6
Last commit
2 hours ago

What is it?

DeepChat is a desktop app that brings cloud models such as OpenAI, Gemini, Anthropic and DeepSeek, as well as local Ollama models, into one interface. It runs long-lived agent sessions with project folders, permission modes, tool output and resumable context, and records session history so work can be recovered and inspected.

The app supports the Model Context Protocol (MCP) with Resources, Prompts and Tools across StreamableHTTP, SSE and Stdio transports. Reusable Skills can be installed from folders, ZIP files or URLs and enabled per conversation. ACP-compatible coding agents can be selected alongside regular models in the model selector.

DeepChat can also be controlled from Telegram, Feishu/Lark, QQBot, Discord and WeChat iLink, letting users switch models, answer pending prompts and stop runs while away from the desktop. It is built with Electron and TypeScript and released under the Apache-2.0 license for Windows, macOS and Linux.

Key features

  • 01Runs cloud and local Ollama models from one desktop interface
  • 02Supports MCP Resources, Prompts and Tools over multiple transports
  • 03Installs and enables reusable Skills per conversation
  • 04Runs ACP-compatible agents as selectable entries in the model selector
  • 05Controls sessions remotely from Telegram, Feishu/Lark, Discord and other messaging apps

Use cases

  • Using multiple LLM providers without switching between separate apps
  • Running coding agents alongside chat models in one local client
  • Monitoring or steering a long-running agent session from a phone via messaging apps
  • Managing local Ollama models through a graphical interface

Quick start

  1. 1

    Install via Homebrew (macOS only)

    brew install --cask deepchat
  2. 2

    Or download the installer for your system from GitHub Releases

  3. 3

    Launch DeepChat, open settings, and add API keys under Model Providers or configure local Ollama

Things to know

  • Homebrew install is available only on macOS; other platforms need the release installers.

Topics

#ai-assistant#llm-client#mcp#electron#ollama#agent#desktop-app#openai#agent-skills#ai

FAQ

What is deepchat?
An open-source, local-first AI agent desktop client supporting MCP, Skills, ACP agents and remote control from messaging apps. DeepChat is a desktop app that brings cloud models such as OpenAI, Gemini, Anthropic and DeepSeek, as well as local Ollama models, into one interface. It runs long-lived agent sessions with project folders, permission modes, tool output and resumable context, and records session history so work can be recovered and inspected.
Is deepchat free to use? What license does it have?
deepchat is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is deepchat still maintained?
Actively maintained. The last commit to deepchat was on Oct 22, 2026.
What can I use deepchat for?
Using multiple LLM providers without switching between separate apps Running coding agents alongside chat models in one local client Monitoring or steering a long-running agent session from a phone via messaging apps Managing local Ollama models through a graphical interface
How do I install deepchat?
Install via Homebrew (macOS only): brew install --cask deepchat Or download the installer for your system from GitHub Releases Launch DeepChat, open settings, and add API keys under Model Providers or configure local Ollama
Where can I get deepchat?
The source code and installation instructions are on GitHub: https://github.com/ThinkInAIXYZ/deepchat

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects