queryverify

QueryVerify is an SQL debug agent that double-checks AI-generated SQL queries by executing them in an isolated sandbox before they reach a production database. It addresses the risk of deploying error-prone queries produced by AI models, allowing verification against test data without affecting live systems. The tool is illustrated through a real-world PostgreSQL example of checking whether foreign key columns ending in _id have indexes across tables.

It begins by generating a tailored PostgreSQL sandbox environment that includes table creation, insertion of diverse test data with both positive and negative cases, and relevant indexes. A visualization of this sandbox is provided to the user. Next it produces an SQL query based on the given prompt, such as one that identifies foreign key columns and reports whether each has an associated index. The generated query is then executed within the sandbox.

If errors occur during execution, users can enable Auto Error Fixing Mode, which attempts automatic resolution while enforcing a safety limit to avoid infinite loops, or supply manual debugging instructions. Upon completion the tool displays the query results, enabling direct comparison between expected outcomes and actual sandbox output to confirm correctness.

The core of the platform consists of an AI agent built on N8N and is available as open source on GitHub. QueryVerify is delivered as a web-based service with documentation, a roadmap, and options to get started or log in. Contact is available at contact@queryverify.com.