AstroLogics API

MCP servers

Give an AI assistant real calculations and our documentation through the Model Context Protocol, with nothing to install. A language model cannot calculate a chart; it can call one and explain the result.

API server: one tool for each public operation, 133 today, with your API key.

https://api.astrologics.in/mcp

Docs server: no key needed. Its tools are search_docs, read_doc, list_docs, list_operations and get_operation.

https://developers.astrologics.in/mcp

Connect in one line

With Claude Code and a test key:

claude mcp add --transport http astrologics https://api.astrologics.in/mcp --header "Authorization: Bearer al_test_YOUR_KEY"

The setup guide also covers the Docs server, Cursor and VS Code.

What each server is for

  • API server: an assistant in your product or workflow can fetch a kundli, a panchang or a match, and answer from the result.
  • Docs server: a coding tool can look up an operation, its parameters and credits while it writes your integration. It needs no key and spends no credits.

Keep the assistant's context small

Listing every operation is a lot for an agent to hold in mind.

  • ?domains=kundli,dasha lists only those groups, using the slugs from the reference.
  • ?compact=1 shortens the tool descriptions, and each call answers without meta.

They work together: https://api.astrologics.in/mcp?domains=vastu&compact=1.

Credits and your data

An MCP call goes through the same checks as a REST call: the same key, rate limits and credits. Only successful calls on a live key are charged, and a test key is never charged. Arguments are treated like any request body and are never written to our database or logs.

What the model does with a result is up to the model and whoever runs it. Our servers do not see its reply, and the wording of its answer is not ours.

A key in an MCP config is a secret key, so keep it out of files you commit.

Next: setup guide · AI assistants · SDKs