Use a real Vedic astrology engine inside any AI

Why this is different: compute, not content

Ask a generic astrology chatbot about your Saturn and it may invent a placement that was never calculated — confident prose over unstable facts. Eternal Evals separates the two jobs. A deterministic engine (Swiss Ephemeris DE441, Lahiri ayanamsa) computes your chart first — planets, houses, 16 divisional charts, dashas, shadbala, yogas — and only then does the AI read what the engine produced. The model never fabricates a degree or a date; it reads them.

The bridge that makes this possible in modern AI apps is MCP — the Model Context Protocol, an open standard for giving assistants callable tools. Our MCP server exposes three: compute_chart, read_section and list_sections.

Fastest path — add the connector to Claude

No install. One URL.

1. In Claude, open Settings → Connectors → Add custom connector.
2. Paste the URL and give it a name:

https://mcp.eternalevals.com/mcp

3. Save. Claude now has compute_chart, read_section and list_sections. Ask it: "Compute my Vedic chart for 7 July 1990, 05:00, Panipat, India — am I Manglik?" and it will call the engine, read the relevant sections, and answer from real numbers.

Custom connectors require a paid Claude plan. On the free plan, use the local server or the ChatGPT GPT below.

Cursor

Add the remote server to Cursor's MCP config (~/.cursor/mcp.json):

{
  "mcpServers": {
    "eternal-evals": { "url": "https://mcp.eternalevals.com/mcp" }
  }
}

Reload Cursor and the three chart tools appear. The same JSON works in any client that supports remote (HTTP) MCP servers.

Run it locally (Claude Desktop, Cline, offline)

Prefer a local process, or your client only supports stdio servers? Download the server and point your client at it.

curl -O https://eternalevals.com/mcp/eternalevals_mcp.py
pip install mcp httpx

Then, for Claude Desktop, add to claude_desktop_config.json (Settings → Developer → Edit Config):

{
  "mcpServers": {
    "eternal-evals": {
      "command": "python",
      "args": ["/absolute/path/to/eternalevals_mcp.py"]
    }
  }
}

Restart the app and the tools appear. It runs on your machine and calls the engine over HTTPS.

No MCP client? Use ChatGPT or the API

Two more doors into the same engine:

ChatGPT — a ready-made Custom GPT that calls the engine through a GPT Action. Zero setup, just chat: Open the ChatGPT GPT →

REST API — for your own app or script. Compute a chart, then read the sections you need:

curl -s -X POST https://eternalevals.com/api/v1/chart \
  -H "Content-Type: application/json" \
  -d '{"date":"1990-07-07","time":"05:00","place":"Panipat, India"}'
# -> { "chart_id": "...", "section_count": 71, "sections": [ ... ] }

Full docs and the OpenAPI schema are on the developers page.

What you can ask, and what the tools do

Once connected, the assistant has three tools:

• compute_chart — give a birth date, time and place; get a chart ID and the list of ~70 readable sections.
• read_section — read any section (or a nested value like houses.10 or planets.Saturn): lagna, planets, houses, vimshottari_dasha, shadbala, yogas, nakshatra, ashtakavarga…
• list_sections — see everything available for a computed chart.

Good questions: "Which Mahadasha am I running and when does it change?" · "Am I Manglik — check from Lagna, Moon and Venus." · "What are my strongest yogas?" · "Is this a Sade Sati phase?" Because every answer is grounded in a computed value, you can keep drilling: ask which chart factor supports a claim and the assistant can point to the exact placement, strength or period.