← Bhasini

Bhasini MCP Server

Give your AI assistant exact spelling in all 22 scheduled languages of India. Bhasini's MCP server transliterates romanised text, converts between Indian scripts, and looks up keyboard maps with the same engine as this website. Connect to the hosted server with one URL, or run it on your own computer.

Tools

ToolWhat it doesExample
transliterateRomanised text to native script, by soundvaNakkam + ta → வணக்கம்
convert_scriptOne Indian script to anotherनमस्ते + kn → ನಮಸ್ತೇ
get_keymapPhonetic keyboard map for a languageKeys for every letter of Meetei Mayek
native_digitsDigits in native numerals2026 + hi → २०२६
list_languagesSupported languages and codeshi, ta, sat, mni, ks, …

Connect (nothing to install)

Server URL: https://mcp.bhasini.com/mcp. No account or API key needed.

Claude (claude.ai and Claude Desktop)

Add Bhasini from Claude's directory →

Or add it by hand: Settings → Connectors → Add custom connector, then paste the URL above.

Claude Code

claude mcp add --transport http bhasini https://mcp.bhasini.com/mcp

ChatGPT

Turn on developer mode in ChatGPT's settings, then add a custom app with the URL above.

OpenAI Codex (~/.codex/config.toml)

[mcp_servers.bhasini]
url = "https://mcp.bhasini.com/mcp"

VS Code (.vscode/mcp.json)

{
  "servers": {
    "bhasini": { "type": "http", "url": "https://mcp.bhasini.com/mcp" }
  }
}

Cursor and other JSON configs

{
  "mcpServers": {
    "bhasini": { "url": "https://mcp.bhasini.com/mcp" }
  }
}

Or run it locally (offline)

You need the .NET SDK. The dnx command comes with .NET 10 or later.

Claude Code

claude mcp add bhasini -- dnx Bhasini.Mcp@0.1.0 --yes

Claude Desktop, Cursor, Windsurf and other JSON configs

{
  "mcpServers": {
    "bhasini": {
      "command": "dnx",
      "args": ["Bhasini.Mcp@0.1.0", "--yes"]
    }
  }
}

OpenAI Codex CLI (~/.codex/config.toml)

[mcp_servers.bhasini]
command = "dnx"
args = ["Bhasini.Mcp@0.1.0", "--yes"]

VS Code (.vscode/mcp.json)

{
  "servers": {
    "bhasini": { "type": "stdio", "command": "dnx", "args": ["Bhasini.Mcp@0.1.0", "--yes"] }
  }
}

With .NET 8 or 9

dotnet tool install -g Bhasini.Mcp

Then use bhasini-mcp as the command, with no arguments.

Try asking

Privacy

Local server: collects nothing, stores nothing and sends nothing. Text is processed in memory on your computer and discarded when each call returns.

Hosted server (mcp.bhasini.com): runs on Cloudflare Workers. The text in each tool call is processed in memory and is not stored or logged. Cloudflare handles the connection, including your IP address, which is used only to limit each client to 120 requests a minute. Request metadata such as time and status may appear in Cloudflare's operational logs.

Your AI assistant handles the conversation under its own privacy policy. See the privacy policy and terms of service.

Source

Open source under the MIT license: github.com/Codegres-com/Bhasini. Questions: naveen@codegres.com.