T20 and ODI Ball by Ball Are Now on Both Engines, Postgres and DuckDB. Run Your SQL Directly on My Server.
Published: August 19, 2026
Both T20 & ODI cricket ball by ball now there on both engines - Postgres & DuckDB. Run your SQL directly on my server. Free, no auth, no key. Calculate strike rates by batter against pace and against spin, over by over run rates through the death overs, head to head records between two teams across a decade, dismissal patterns by bowler type, home and away splits by venue... etc .. etc
Auto-refreshes twice a day.
Tip: most calls come in over GET, and the commonest error is still the URL rather than the SQL, so if a query comes back looking mangled, send it as a POST with a JSON body instead.
Docs
- for humans: db-mcp.tigzig.com/v1/redoc
- for your AI: db-mcp.tigzig.com/v1/openapi.json
Credit: data sourced from Cricsheet, ODC-BY 1.0.
No MCP required, just ask your AI worker
If you have a ChatGPT or Claude subscription - both come with a Cowork app - you can simply share the message / link above and ask it to explore the data for you. Also with a Gemini subscription which comes with Antigravity .. which is actually a full fledged AI Coder. Comes with the $20 / Rs. 1,950 subscription also.
No MCP required.. just ask like in a regular chat bot ... they will go and fetch the data directly via API .. copy it to your local drive, create excel, deck, reports for you also if you'd like.
Just download their desktop app and login .. that's it ..
For more on using AI Workers, Tigzig AI Worker hub at tigzig.com/claude-for-analytics
What is allowed, and what gets refused
It is read only. Writes, DDL, and anything that reads a file or a URL are refused. There are caps on functions that can generate rows from nothing or spin CPU without limit, and the refusal names the specific function so you can drop that one and retry rather than rewriting the query.
Schema discovery is deliberately open. information_schema works, SHOW TABLES and DESCRIBE work, so you or your agent can find the columns without me publishing them anywhere.
A couple of the guards were tighter than they needed to be in the first week and got loosened once I realized I was refusing many regular queries.
The Postgres now runs on my self hosted box, so way faster than earlier which was on a managed instance.