Tutorials

Jupyter notebooks that call MANNA’s tools directly from Python through fastmcp’s in-memory client, so they run with only pip install manna-mcp and network access to the archives. No LLM is involved; the point is to see exactly what each tool returns.

from fastmcp import Client
from manna.app import build_mcp

client = Client(build_mcp())
async with client:
    result = await client.call_tool("list_archives", {})
    print(result.structured_content["count"])

  1. 1. First query — connect, list archives, resolve a name, describe a table, run a small query, read the inline envelope.

  2. 2. Large results — an oversize query promoted to an async job; poll, fetch the result with pyvo, save it with the recipe.

  3. 3. Images and catalogs — the workflow tools: find and display an image, cone search, count, survey.

  4. 4. Archive notes and errors — preview a table, a classic Data Lab mistake, the error envelope, the up-front note, truncated=true.

Recorded against MANNA 0.11.0 on 2026-10-08. The outputs are committed; the site never re-runs them. To re-record, see Development.