Unichat

Provides an MCP server to route Unichat requests through multiple AI models using predefined prompts.

python

GitHub Stars

python

Language

7 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "mcp-mirror-amidabuddha_unichat-mcp-server": {
      "command": "uv",
      "args": [
        "--directory",
        "{{your source code local directory}}/unichat-mcp-server",
        "run",
        "unichat-mcp-server"
      ],
      "env": {
        "UNICHAT_MODEL": "SELECTED_UNICHAT_MODEL",
        "UNICHAT_API_KEY": "YOUR_UNICHAT_API_KEY"
      }
    }
  }
}

You can run the Unichat MCP Server to connect to OpenAI and other AI vendors through the MCP protocol. It exposes a single tool to send conversations and supports predefined prompts to review, document, explain, or rework code. This server is available in Python and can be run locally or deployed to respond to MCP client requests with a consistent interface.

How to use

Start the server in stdio mode and connect an MCP client to send messages to the unichat tool. Use the predefined prompts to perform actions like code reviews, documentation generation, explanations, and code reworks. The server expects a model and API key for the chosen vendor, supplied via environment variables. When running, you will receive structured responses that you can surface in your application or tooling.

How to install

Prerequisites: you need a runtime that supports MCP servers (the runtime in this case is provided by the tooling used for MCP servers). You should have a directory containing the server source code and the runtime available in your environment.

Additional configuration and notes

The server configuration supports two MCP endpoints in stdio mode for development and production publishing. Each endpoint runs as a separate stdio process and uses environment variables to supply the selected model and your vendor API key.

{
  "mcpServers": {
    "unichat_mcp_dev": {
      "command": "uv",
      "args": [
        "--directory",
        "{{your source code local directory}}/unichat-mcp-server",
        "run",
        "unichat-mcp-server"
      ],
      "env": {
        "UNICHAT_MODEL": "SELECTED_UNICHAT_MODEL",
        "UNICHAT_API_KEY": "YOUR_UNICHAT_API_KEY"
      }
    },
    "unichat_mcp_pub": {
      "command": "uvx",
      "args": [
        "unichat-mcp-server"
      ],
      "env": {
        "UNICHAT_MODEL": "SELECTED_UNICHAT_MODEL",
        "UNICHAT_API_KEY": "YOUR_UNICHAT_API_KEY"
      }
    }
  }
}

Notes on tools and prompts

The MCP server exposes a single tool named unichat that sends a request to Unichat. It supports the following prompts to operate on code and text:

Prompts and tools

code_review Review code for best practices, potential issues, and improvements. Arguments: code (string, required). The prompt analyzes the code and returns actionable recommendations.

document_code Generate documentation for code including docstrings and comments. Arguments: code (string, required). The prompt returns structured documentation.

Available tools

unichat

Send a request to Unichat via the MCP tool and receive a response

code_review

Review code for best practices, potential issues, and improvements

document_code

Generate documentation for code including docstrings and comments

explain_code

Explain how a piece of code works in detail

code_rework

Apply requested changes to the provided code

Built by

VeilStrat

AI signals for GTM teams