MCP-CLIENT-BROSWER

by autoexpect

5 332 downloads Not rated yet MIT

About

TypeScript SDK for using large language models with an MCP server directly in the browser, powered by Cherry Studio's prompt templates.

Details

License
MIT

Explore

- 🌐 Runs entirely in the browser (no backend required)
- ⚑ Communicates with MCP-compatible servers
- 🧠 Built-in support for Cherry Studio's MCP prompt format
- πŸ› οΈ Simple, extendable, and TypeScript-friendly API
- πŸ“¦ Designed for easy integration into modern web apps

Install via npm i @autoexpect/mcp-sseclient-browser. Instantiate the MCPClient class with an array of server configurations and an options object (containing base URL, API key, and model), then call connect() and processQueryStream() with system and user prompts, a boolean flag, and callback functions for chunks and the full response.

mcp-client-browser is a TypeScript library that brings MCP (Model Control Protocol) support to the browser. It enables seamless interaction with large language models (LLMs) through an MCP server, leveraging Cherry Studio's prompt templates and communication logic.

✨ Features

- 🌐 Runs entirely in the browser (no backend required)
- ⚑ Communicates with MCP-compatible servers
- 🧠 Built-in support for Cherry Studio's MCP prompt format
- πŸ› οΈ Simple, extendable, and TypeScript-friendly API
- πŸ“¦ Designed for easy integration into modern web apps

πŸš€ Getting Started

Installation

npm i @autoexpect/mcp-sseclient-browser

Usage

let mcpClient = new MCPClient(
  [
    {
      url: new URL('/api/mcps/sse', window.location.origin)
    }
  ],
  {
    baseUrl: window.location.origin + "/v1",
    apiKey: "sk-default",
    model: "deepseek-v3",
  }
);
await mcpClient.connect();
await mcpClient.processQueryStream(
  "System Prompt",
  "User Prompt",
  true,
  (chunk) => {
  },
  (fullResponse) => {
  }
);

> βœ… Make sure your MCP server supports CORS and the Cherry prompt specification.

πŸ’‘ Use Cases

- AI chatbots powered by LLMs directly in the browser
- Embedding customizable prompt workflows into frontend apps
- Creating interactive educational or creative writing tools

🧩 Integrations

- Cherry Studio: Easily plug in Cherry-style prompt templates
- Compatible with most OpenAI-compatible or custom MCP servers

πŸ›‘οΈ Security & Privacy

This library does not store user data. All interactions happen directly between the browser and the configured MCP server.

πŸ“„ License

MIT License

No reviews yet β€” be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.