We said MAYBE on August 13, 2026. Not settled — due August 13, 2027.
Read this with the caveat. We tested the engine that produced this verdict against 292 launches whose outcomes we already knew, and could not show it predicted which survived. Some of its data sources were also dead at the time of scoring. The verdict stays up, dated and unedited, because a record you can quietly revise is not a record — but it is worth less than it looked when it was written.
There's a genuine, defensible open-source wedge in SQL-agent evaluation with real gaps competitors leave open, but demand is inferred from the adjacent agent market rather than directly proven, and eval tooling is a chronically hard thing to charge for. This is a plausible BUILD as an open-source-led wedge, but with real monetization risk — treat it as a lean experiment, not a conviction bet.
Multiple funded/commercial NL-to-SQL players prove there's real money in the SQL agent space, and HN discussions draw strong engagement — but demand for the *evaluation layer* specifically is inferred, not directly evidenced, with zero Reddit posts and no strong keywords.
A clear open-source wedge exists — no incumbent offers a standardized eval framework with pre-built correctness/latency/cost metrics — and competitors are either generic frameworks or enterprise-focused, both beatable by an API-first indie tool. Deducted because the buyer (data/ML engineers) evaluates SQL agents only after adopting one, making this a dependent, second-order market.
Evaluation tooling is notoriously hard to monetize as a standalone product — engineers happily use open-source eval harnesses for free and rarely pay, so an open framework may win GitHub stars while generating near-zero revenue. The demand signal here is entirely borrowed from the agent-building market, not the eval niche itself, and this could easily be a feature swallowed by LangSmith or the agent vendors themselves.
What is worth more than this page. The register records what became of 6,266 real launches. No engine has to be right for that to be true.