We present SpotIt+, an open-source tool for evaluating Text-to-SQL systems via bounded equivalence verification. Given a generated SQL query and the ground truth, SpotIt+ actively searches for database instances that differentiate the two queries. To ensure that the generated counterexamples reflect practically relevant discrepancies, we introduce a best-effort constraint-mining pipeline that combines rule-based specification mining with LLM-based validation over example databases. Experimental results on the BIRD dataset show that the mined constraints enable SpotIt+ to generate more realistic differentiating databases, while preserving its ability to efficiently uncover numerous discrepancies between generated and gold SQL queries that are missed by standard test-based evaluation.
@article{arxiv.2603.04334,
title = {SpotIt+: Verification-based Text-to-SQL Evaluation with Database Constraints},
author = {Andrew Tremante and Yang He and Rocky Klopfenstein and Yuepeng Wang and Nina Narodytska and Haoze Wu},
journal= {arXiv preprint arXiv:2603.04334},
year = {2026}
}