Vedic Astrology MCP Server (Kundli, Panchang) by DivineAPI

The official DivineAPI MCP server for Vedic (Indian) astrology. It lets AI assistants such as Claude, ChatGPT and VS Code agents fetch Panchang, kundli, dasha, dosha and matchmaking data from DivineAPI. Use the hosted server at mcp.divineapi.com, or run this one locally with your DivineAPI API key. Vedic text is available in 8 Indian languages: English (en), Hindi (hi), Bengali (bn), Marathi (ma), Tamil (tm), Telugu (tl), Malayalam (ml) and Kannada (kn). Calculations use the Lahiri ayanamsa. 14-day free trial (credit card required to activate the trial): https://divineapi.com/start-trial

Tools also cover Lal Kitab, Jaimini and Hindu festival dates, returned as JSON (and SVG for charts). Product page: Vedic astrology API.

Hosted server URL:

https://mcp.divineapi.com/indian/mcp

One-click install (adds the hosted server with placeholder credentials):

After installing, replace YOUR_API_KEY and YOUR_AUTH_TOKEN in the server's headers with your own values. In VS Code you can use the inputs version of the config below instead, so the values are stored securely.

Connect in 1 minute

1. Get your credentials. Start the 14-day free trial (credit card required to activate the trial), then copy your API key and auth token from the DivineAPI dashboard.

2. Add the server to your client. The hosted server accepts your credentials in one of two ways:

• Request headers: X-Divine-Api-Key and X-Divine-Auth-Token (Cursor, VS Code, Claude Code).
• Sign-in page (OAuth): the client opens a DivineAPI page where you paste the API key and auth token once (Claude.ai and Claude Desktop custom connectors).

Claude.ai and Claude Desktop (custom connector)

Custom connectors added in Claude.ai also appear in Claude Desktop.

1. Open Customize > Connectors, click + Add, then Add custom connector.
2. Name: DivineAPI Vedic Astrology. URL: https://mcp.divineapi.com/indian/mcp. Click Continue.
3. Keep the detected sign-in settings and click Add. When you connect, a DivineAPI sign-in page opens: paste your API key and auth token and click Connect.

If your Claude plan offers request headers for custom connectors, you can instead choose no sign-in and add the two headers X-Divine-Api-Key and X-Divine-Auth-Token.

Team and Enterprise plans: an owner adds the connector under Organization settings > Connectors first. Steps from Claude's custom connector guide.

Cursor

~/.cursor/mcp.json (all projects) or .cursor/mcp.json (one project):

{
  "mcpServers": {
    "divineapi-indian-astrology": {
      "url": "https://mcp.divineapi.com/indian/mcp",
      "headers": {
        "X-Divine-Api-Key": "YOUR_API_KEY",
        "X-Divine-Auth-Token": "YOUR_AUTH_TOKEN"
      }
    }
  }
}

VS Code (GitHub Copilot agent mode)

.vscode/mcp.json. VS Code asks for both values on first start and stores them securely:

{
  "inputs": [
    { "type": "promptString", "id": "divine-api-key", "description": "DivineAPI API key", "password": true },
    { "type": "promptString", "id": "divine-auth-token", "description": "DivineAPI auth token", "password": true }
  ],
  "servers": {
    "divineapi-indian-astrology": {
      "type": "http",
      "url": "https://mcp.divineapi.com/indian/mcp",
      "headers": {
        "X-Divine-Api-Key": "${input:divine-api-key}",
        "X-Divine-Auth-Token": "${input:divine-auth-token}"
      }
    }
  }
}

Claude Code

claude mcp add --transport http divineapi-indian-astrology https://mcp.divineapi.com/indian/mcp \
  --header "X-Divine-Api-Key: YOUR_API_KEY" \
  --header "X-Divine-Auth-Token: YOUR_AUTH_TOKEN"

Run /mcp inside Claude Code to check the connection.

Clients that send only one credential field

If a client can send only a single bearer token and no custom headers, send Authorization: Bearer YOUR_API_KEY:YOUR_AUTH_TOKEN (API key, a colon, then the auth token). The server splits it into the two credentials.

Example prompts

• "What is today's Panchang for New Delhi? Include tithi, nakshatra, yoga, karana and Rahu Kaal."
• "I was born on 15 March 1990 at 14:30 in Mumbai. Show my D1 and D9 charts and my current Vimshottari dasha."
• "Am I Manglik? Is Sadhe Sati active for me right now?"
• "Match these two birth charts with Ashtakoot Milan and explain the guna score."
• "List the Hindu festivals in Kartik month this year, with dates for Delhi."
• "Give me my Lal Kitab planetary positions and the debts (rin) in my chart."

Tool groups

The table groups the tools defined in server.py. The hosted server is updated more often and can offer more tools; your MCP client always shows the live list.

All tools are read-only (readOnlyHint: true) and each tool call makes one request to the DivineAPI REST API on the astroapi-1, astroapi-2, astroapi-3 or astroapi-8 host.

Inputs the model needs to know

• Dates are separate day, month, year fields (plus hour, min, sec for birth charts; send sec as 0 if unknown).
• place is a plain lowercase city string; lat and lon are decimals.
• tzone is a decimal UTC offset, for example 5.5 for India. Never +5:30.
• Calculations are sidereal with the Lahiri ayanamsa.
• lan sets the response language. Vedic responses support 8 Indian languages with DivineAPI's own codes:

Self-hosting

You only need this if you want to run the server yourself instead of using the hosted URL. The server reads your credentials from two environment variables: DIVINE_API_KEY and DIVINE_AUTH_TOKEN.

Local (stdio) with uv or pip

# uv (no install step)
uvx divineapi-indian-astrology-mcp

# or pip
pip install divineapi-indian-astrology-mcp
divineapi-indian-astrology-mcp

Claude Desktop local config (claude_desktop_config.json; on macOS ~/Library/Application Support/Claude/, on Windows %APPDATA%\Claude\):

{
  "mcpServers": {
    "divineapi-indian-astrology": {
      "command": "uvx",
      "args": ["divineapi-indian-astrology-mcp"],
      "env": {
        "DIVINE_API_KEY": "YOUR_API_KEY",
        "DIVINE_AUTH_TOKEN": "YOUR_AUTH_TOKEN"
      }
    }
  }
}

The PyPI package can be newer than the code in this repository. To run exactly this repository's code:

git clone https://github.com/DivineAPI/mcp-indian-astrology.git
cd mcp-indian-astrology
pip install .
DIVINE_API_KEY=YOUR_API_KEY DIVINE_AUTH_TOKEN=YOUR_AUTH_TOKEN divineapi-indian-astrology-mcp

Remote (streamable HTTP) with Docker

docker build -t divineapi-indian-mcp .
docker run -p 8000:8000 -e MCP_JWT_SECRET=a-long-random-string divineapi-indian-mcp

The MCP endpoint is then http://localhost:8000/mcp. Clients send their own X-Divine-Api-Key and X-Divine-Auth-Token headers. docker compose up maps the same server to port 8001.

The OAuth sign-in flow expects HTTPS on MCP_HOST behind a reverse proxy that serves the server under /indian/ (see nginx.conf). Header authentication works without it.

Which plan includes this

The tools call the Vedic endpoints, so they use your Vedic plan (Vedic Sampoorna, Vedic Ananta or Vedic Prakash); a tool works when your plan includes its endpoint. Compare plans at divineapi.com/pricing.

FAQ

Do I need a DivineAPI account to use the Vedic tools? Yes. Each panchang, kundli or matching tool sends your API key and auth token to the Vedic REST endpoints. Start with the 14-day free trial (credit card required to activate the trial); plans and prices are at divineapi.com/pricing.

Can Claude Desktop or Cursor ask for a kundli directly? Yes. Any client that supports remote MCP over streamable HTTP (Claude.ai, Claude Desktop, Claude Code, Cursor, VS Code) can call the kundli and panchang tools from a plain-language prompt. A client that only starts local servers can run divineapi-indian-astrology-mcp over stdio (see Self-hosting).

Why does my client show Vedic tools that are not in the table? The table follows this repository's server.py. The hosted server and the PyPI package can be newer than this code, and your client always shows the live list.

Which ayanamsa and zodiac does it use? Sidereal zodiac with the Lahiri ayanamsa, calculated with Swiss Ephemeris.

Can it draw charts? Yes. Chart tools (divine_get_horoscope_chart, divine_get_sub_planet_chart, the Lal Kitab and matching chart tools) return an SVG plus a base64 image.

I want the REST API, not MCP. Call the same Vedic endpoints over plain HTTP with the kundli-api and panchang-api quickstarts or the API reference. SDKs: pip install divineapi, npm install divineapi, composer require divineapi/divineapi.

Related repos

Support

• MCP overview: divineapi.com/mcp
• API reference: developers.divineapi.com/indian-api
• Help center: support.divineapi.com
• API status: status.divineapi.com

License

MIT. See LICENSE.