Hacker News front pageJul 12, 2026, 04:03 UTC+8

Sqlsure: deterministic semantic checks for AI-generated SQL, catching fan-out double-counting and wrong join keys before they hit production

Show HN: Sqlsure – deterministic semantic checks for AI-generated SQL

Sqlsure is a deterministic semantic checker for AI-generated SQL. It catches fan-out double-counting, additivity violations, wrong join keys, and policy breaches before the query runs. The tool found real bugs in the BIRD and Spider text-to-SQL benchmarks, which suggests those benchmarks still miss a fair number of semantic errors. The post does not disclose detection rates, false-positive rates, or performance overhead.

Why it matters: Clear positioning: deterministic rules catching probabilistic model mistakes, validated on authoritative benchmarks. But it's a niche tool with narrow audience, missing R axis. Scored 72 at the featured threshold.

Read the original ↗Export Markdown

Take

I clicked on this because it hits a real pain point: AI-generated SQL often parses fine but produces logic errors at runtime—fan-out double-counting from joins, additivity violations in aggregates, wrong join keys. Sqlsure tackles these with hardcoded rules rather than model inference, which is why they call it deterministic.

The author ran it against BIRD and Spider, two standard text-to-SQL benchmarks, and found semantic bugs the benchmarks missed. That's not surprising—BIRD and Spider mainly check if execution results match, so they tolerate intermediate logic errors.

But the README doesn't disclose two numbers I'd want: what percentage of errors it catches, and how often it flags correct queries as wrong. Without those, it's hard to know if this is quiet enough for a CI pipeline. No performance overhead data either. I'd treat it as a lint tool for specific SQL anti-patterns rather than a general semantic validator.

Source: Hacker News front page · github.com
Published: Jul 12, 2026, 04:03 UTC+8 · Jul 12, 2026