mcpdata Request an invite

Answers / Problems and fixes

My AI only analysed the first few thousand rows. Why?

Recipe verified on Ran on the public demo (read-only)
Short answer

Raw rows come back in pages, and an assistant that reads rows instead of asking for totals sees only the first page. Ask for the summary instead (group and aggregate in the query) and check the truncated flag: it says when a result was cut short.

Steps

  1. Connect your assistant to the public demo. It's a read-only instance with 10 million synthetic orders, so you can try this before you have your own. In Claude Code: 
    claude mcp add --transport http mcpdata-demo https://demo.mcpdata.org/mcp --header "Authorization: Bearer cdb_live_PkVoldus8qcE-hqffjajNZzQwiA7rTHMoR4RLSAxsCE"

    The key is public on purpose: it can only read the demo's sample data. For Claude Desktop, Cursor and other apps, see Connect your AI.

  2. Ask your question. Give your AI this prompt:
    Using mcpdata, how many orders per region are in Demo.Orders? Don't read the raw rows.
What your AI does: the tool calls it makes for this prompt, and what came back
query(table_name="Demo.Orders", columns=["order_id", "region", "revenue"]) → 1,000 rows (more available)
query(table_name="Demo.Orders", group_by=["region"], agg=[{"func": "count", "column": "order_id", "alias": "orders"}]) → 3 rows
What you should get · the recipe's real result from its last run
regionorders
APAC2,008,864
Americas4,194,772
EMEA3,801,364

Synthetic data: the demo's orders are generated, not real customers.

Ask Claude with this prompt Ask ChatGPT with this prompt

The assistant needs mcpdata connected (step 1) to run the query.

Why this works

The first call below asks for raw rows: it gets 1,000 of 10 million, marked truncated. The second asks the database to count, and gets the exact answer for every row in three lines. Telling the assistant to aggregate (or reading the server's query guide, which says the same) avoids the trap.

Related questions