mcp-server-sarvam

MCP Server for Sarvam

A Model Context Protocol (MCP) server that provides access to Sarvam's language processing capabilities. This server enables LLMs to perform tasks such as translation, language identification, and transliteration, specifically supporting translation between English and Indic languages, as well as transliteration across different scripts.

Components

Tools

The server implements the following tools:

• translate_text: Translates text between English and Indic languages.
• identify_language: Detects the language of the input text.
• transliterate_text: Converts text from one script to another while preserving pronunciation.

Configuration

The server can be configured with the following arguments:

• --api-key (required): The API key for authentication with Sarvam services.

Quickstart

Install

Installing via Smithery

To install the MCP Server for Sarvam automatically via Smithery:

npx -y @smithery/cli install mcp-server-sarvam --client claude

Development/Unpublished Servers Configuration

"mcpServers": {
  "sarvam": {
    "command": "uv",
    "args": [
      "--directory",
      "{{PATH_TO_REPO}}",
      "run",
      "mcp-server-sarvam",
      "--api-key",
      "{{SARVAM_API_KEY}}"
    ]
  }
}

Published Servers Configuration

"mcpServers": {
  "sarvam": {
    "command": "uvx",
    "args": [
      "mcp-server-sarvam",
      "--api-key",
      "{{SARVAM_API_KEY}}"
    ]
  }
}

Replace {{PATH_TO_REPO}} and {{SARVAM_API_KEY}} with the appropriate values.

Development

Building and Publishing

To prepare the package for distribution:

uv sync  # Sync dependencies and update lockfile
uv build  # Build package distributions

This will create source and wheel distributions in the dist/ directory.

Publish to PyPI

uv publish

Note: You'll need to set PyPI credentials via environment variables or command flags:

• Token: --token or UV_PUBLISH_TOKEN
• Username/password: --username / UV_PUBLISH_USERNAME and --password / UV_PUBLISH_PASSWORD

Debugging

Since MCP servers run over stdio, debugging can be challenging. For the best debugging experience, we strongly recommend using the MCP Inspector.

You can launch the MCP Inspector via npm with this command:

npx @modelcontextprotocol/inspector uv --directory {{PATH_TO_REPO}} run mcp-server-sarvam

Upon launching, the Inspector will display a URL that you can access in your browser to begin debugging.