ParityDIff
Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

ParityDiff helps engineering teams find and fix breaking changes in third-party APIs before they disrupt their applications. It connects vendor API changes to the customer’s actual code, identifying affected functions and call paths so engineers can see which changes matter to them. I’m building an AI layer that uses this evidence to explain the impact, propose migration patches, and check those patches against explicitly defined behavior. Engineers can review the evidence, provide feedback, and request revisions while the original expectations remain fixed. Coding assistants will also be able to access ParityDiff’s evidence and checks. The goal is to turn API maintenance from manually reading changelogs and debugging broken integrations into reviewing evidence-backed, validated fixes.