ErrLookup turns library source code into a searchable database of every error message

Error messages are often the first — and worst — documentation a developer encounters. They tell you something is broken, but rarely why, and even less often how to fix it. ErrLookup, a project from Standard Beagle, is trying to change that by systematically crawling open-source libraries and documenting every user-facing error it can find.

The site, which launched recently, covers 296 repositories, with the most-documented libraries listed first. For each error, ErrLookup provides the exact message, the likely cause, a list of fixes in order, and defensive patterns to avoid triggering it in the first place. The goal, according to the project, is to serve three audiences: coding agents that need to resolve errors autonomously, answer engines that field developer questions, and humans who are just trying to ship.

One of the more interesting aspects is the emphasis on provenance. Every record is AI-assisted analysis of the cited commit, pinned to its SHA. That means you can trace any documented error back to the exact version of the code it came from, which is a welcome step beyond the usual blog-post-style error roundups that go stale quickly. The project also makes its dataset contract public, and invites users to request crawls for libraries that aren't yet covered.

For developers, the immediate utility is clear: instead of pasting an error into a search engine and wading through forum threads, you can look it up directly in a structured, version-aware index. For AI coding tools, the potential is even bigger — a reliable, structured source of error knowledge could be the difference between an agent that flails and one that fixes issues on the first try.

The project is open source and looking for GitHub stars to improve discoverability. It's early days, but if the error database grows and stays accurate, ErrLookup could become a standard reference for both human and machine developers.