https://github.com/bigsweetpotatostudio/hyperchat
HyperChat is a Chat client that strives for openness, utilizing APIs from various LLMs to achieve the best Chat experience, as well as implementing productivity tools through the MCP protocol.
https://github.com/bigsweetpotatostudio/hyperchat
agent chat-application llm local-agent mcp modelcontextprotocol
Last synced: 8 months ago
JSON representation
HyperChat is a Chat client that strives for openness, utilizing APIs from various LLMs to achieve the best Chat experience, as well as implementing productivity tools through the MCP protocol.
- Host: GitHub
- URL: https://github.com/bigsweetpotatostudio/hyperchat
- Owner: BigSweetPotatoStudio
- License: other
- Created: 2024-12-14T06:46:41.000Z (about 1 year ago)
- Default Branch: alpha
- Last Pushed: 2025-05-11T16:13:20.000Z (8 months ago)
- Last Synced: 2025-05-14T23:06:17.815Z (8 months ago)
- Topics: agent, chat-application, llm, local-agent, mcp, modelcontextprotocol
- Language: TypeScript
- Homepage:
- Size: 11.8 MB
- Stars: 570
- Watchers: 8
- Forks: 57
- Open Issues: 15
-
Metadata Files:
- Readme: README.md
- Changelog: ChangeLog.md
- License: LICENSE
Awesome Lists containing this project
README
[中文](README.zh.md) | [English](README.md)
## Introduction
HyperChat is an open-source Chat client that supports MCP and can utilize the APIs of various LLMs to achieve the best Chat experience and productivity tools.
[](https://github.com/BigSweetPotatoStudio/HyperChat/actions/workflows/build.yml)
[](https://www.npmjs.com/package/@dadigua/hyper-chat)
[](https://npm-stat.com/charts.html?package=@dadigua/hyper-chat)
* Supports OpenAI style LLMs: `OpenAI`, `Claude(OpenRouter)`, `Qwen`, `Deepseek`, `GLM`, `Ollama`.
* Fully supports MCP.
## DEMO
* [HyperChat](https://hyperchat.dadigua.men/123456/) on Docker
## Features:
- [x] **🪟Windows** + **🍏MacOS** + **Linux**
- [x] Command line operation, `npx -y @dadigua/hyper-chat`, default port 16100, password 123456, Web access http://localhost:16100/123456/
- [x] Docker
* Command line version `docker pull dadigua/hyperchat-mini:latest`
* Ubuntu desktop + Chrome + BrowserUse version (**coming soon**)
- [x] `WebDAV` synchronization
- [x] `MCP` extensions
- [x] Supports dark mode🌙
- [x] Resources, Prompts, Tools support
- [x] Supports both English and Chinese
- [x] Supports `Artifacts`, `SVG`, `HTML`, `Mermaid` rendering
- [x] Supports defining Agents, can preset prompts, select permitted MCPs
- [x] Supports scheduled tasks, designating Agents to complete tasks at specified times and view task completion status.
- [x] Supports `KaTeX`, displaying mathematical formulas, code rendering with highlights and quick copy
- [x] Added `RAG`, based on the MCP knowledge base
- [x] Introduced ChatSpace concept, supports multi-conversation chatting
- [x] Supports model comparison in chat
### TODO:
- Support official Claude protocol
### LLM
| LLM | Usability | Remarks |
| -------- | ------ | -------------------------- |
| claude | ⭐⭐⭐⭐⭐ | No explanation |
| openai | ⭐⭐⭐⭐ | Also perfectly supports multi-step function calls (gpt-4o-mini also works) |
| gemini flash 2.0 | ⭐⭐⭐⭐ | Very useful |
| qwen | ⭐⭐⭐⭐ | Very useful |
| doubao | ⭐⭐⭐ | Feels okay to use |
| deepseek | ⭐⭐⭐ | Multi-step function calls may have issues |
## Usage
* 1. Configure APIKEY, ensure that your LLM service is compatible with OpenAI style.
* 2. Ensure that `uv + nodejs` and others are installed on your system.
### [uvx & uv](https://github.com/astral-sh/uv)
Install via command line, or check the official Github tutorial [uv](https://github.com/astral-sh/uv)
```
# MacOS
brew install uv
# windows
winget install --id=astral-sh.uv -e
```
### [npx & nodejs](https://nodejs.org/en)
Install via command line, or download from the official site, official [nodejs](https://nodejs.org/en)
```
# MacOS
brew install node
# windows
winget install OpenJS.NodeJS.LTS
```
## Development
```
cd electron && npm install
cd web && npm install
npm install
npm run dev
```
## Telegram
[HyperChat User Communication](https://t.me/dadigua001)
#### Chat supports model comparison
#### Supports clicking tool name for direct invocation and debugging
#### MCP tool prompts + Dynamic modification of LLM tool parameters
#### Supports @ for quick input + Invoke Agent
#### Supports `Artifacts`, `SVG`, `HTML`, `Mermaid` rendering,
#### Supports selecting MCP + selecting individual Tools
#### You can access from anywhere + any device via the Web, and set a password
#### Invokes terminal MCP to automatically analyze ASAR files + help unpack
#### Invokes terminal to view interface
#### Gaode Map MCP
#### One-click to write a webpage and publish to (Cloudflare)
#### Invokes Google search, asking what the TGA Game of the Year is
#### What are the limited-time free games? Please visit the website, invoke the tool
#### Opens the webpage, analyzes the results, and writes to a file
#### Through web tools + command line tools, opens Github README to learn + GIT clone + sets up the development environment
#### Multi-chat Workspace + Night Mode
#### Scheduled task list + Scheduled message sending to Agent to complete tasks
#### Install MCP from third parties (supports any MCP)
#### H5 interface
#### Test the model's capabilities
#### Knowledge Base
## Disclaimer
* This project is for learning and communication purposes only. If you use this project for any operation, such as web scraping, it is unrelated to the developers of this project.