Repolode
Submit
BigSweetPotatoStudio

BigSweetPotatoStudio/HyperChat

A local-first AI chat client that connects to multiple LLM providers and adds productivity tools through the MCP protocol.

Open-source alternative to:ChatGPTClaude.ai

View on GitHub
Stars
709
Forks
74
Open issues
30
Last commit
2 months ago

What is it?

HyperChat is a desktop and command-line chat client for working with large language models. It connects to several LLM providers through their APIs, and it stores agent configuration, memory and chat history as files inside a .hyperchat/ directory, so the setup can live alongside a project and be versioned with Git.

Agents are defined in YAML with a model, a system prompt and a list of allowed MCP servers. Each agent can also keep its own memory file and a set of Markdown command templates, which are invoked with slash commands such as /review or /bug-fix and can reference files with @path syntax. Tools are provided through the Model Context Protocol, letting agents work with local files and other MCP-compatible services.

The project offers two interfaces: a web UI for managing multiple workspaces in tabs with shared MCP servers, and a CLI that starts a named agent directly for one-off prompts or interactive chat. Configuration is read from environment variables and .env files at several levels, with CLI flags taking the highest priority.

Key features

  • 01Connects to multiple LLM providers such as OpenAI, Claude, Gemini, Kimi and Qwen.
  • 02Extends agents with tools via the Model Context Protocol.
  • 03Stores agent definitions, memory and chat logs as files in a project-level .hyperchat/ directory.
  • 04Provides custom slash commands defined as Markdown templates.
  • 05Supports @file references to include multiple files in a prompt.
  • 06Offers both a multi-workspace web interface and a CLI for direct agent use.

Use cases

  • Keeping a project-specific AI assistant whose configuration and memory are committed alongside the code.
  • Running a named agent from scripts or CI jobs with a one-line command.
  • Chatting with several LLM providers from one client while giving the model access to local files through MCP tools.

Quick start

  1. 1

    Install the CLI globally from npm

    npm install -g @dadigua/hyperchat
  2. 2

    Set your API key and default model via environment variables

    export HyperChat_API_KEY=your-api-key
  3. 3

    Send a first message to the default model

    hyperchat "你好，世界！"

Things to know

  • The README's license field is unrecognized, so check the LICENSE file before commercial use.
  • Setup requires configuring an API key for an LLM provider.
  • The 2.0 version is in migration and the README describes it as Vibe Coded, so stability may vary.

Topics

#llm#mcp#ai-agent#chat-client#typescript#local-first#developer-tools#agent#chat-application#local-agent

FAQ

What is HyperChat?
A local-first AI chat client that connects to multiple LLM providers and adds productivity tools through the MCP protocol. HyperChat is a desktop and command-line chat client for working with large language models. It connects to several LLM providers through their APIs, and it stores agent configuration, memory and chat history as files inside a .hyperchat/ directory, so the setup can live alongside a project and be versioned with Git.
Is HyperChat free to use? What license does it have?
HyperChat is open source under the Other license. Check the license terms before using it in commercial products.
Is HyperChat still maintained?
Updated within 6 months. The last commit to HyperChat was on Aug 26, 2026.
What can I use HyperChat for?
Keeping a project-specific AI assistant whose configuration and memory are committed alongside the code. Running a named agent from scripts or CI jobs with a one-line command. Chatting with several LLM providers from one client while giving the model access to local files through MCP tools.
How do I install HyperChat?
Install the CLI globally from npm: npm install -g @dadigua/hyperchat Set your API key and default model via environment variables: export HyperChat_API_KEY=your-api-key Send a first message to the default model: hyperchat "你好，世界！"
Where can I get HyperChat?
The source code and installation instructions are on GitHub: https://github.com/BigSweetPotatoStudio/HyperChat

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

Related projects