SpecDrift Scanner · Open Source

Your code works. But does it do what was actually requested?

Product teams waste up to 70% of rework time on code that silently drifts from specs. SpecDrift Scanner scans your completed issues against merged PRs and shows you exactly where implementation diverged from intent, with evidence.

View on GitHub
Try online for freecoming soon
Live example

See what the scanner finds

Real scan of Wagtail, a popular open-source Django CMS. 30 days, 20 issue-PR pairs analyzed.

SpecDrift Scanner report — real scan of Wagtail
Concept

What is Intent Drift?

BusinessrequirementsSpecs &ticketsCode(merged PR)drift accumulates heredrift accumulates here

At every handoff, from business goal to spec and from spec to code, small misalignments creep in. The code compiles. Tests pass. The PR gets approved. But the implementation quietly deviates from what was actually requested.

Existing review bots check code quality: syntax, style, security patterns. They don't use the full context of what was requested to evaluate whether the code got it right.

The result: the product ships behavior that nobody asked for. Rework follows. Deadlines slip.

Join the waitlist

We're building a full product to solve this

SpecDrift Scanner is the first step. The full Owerview AI platform will monitor intent alignment continuously across your GitHub, Jira, Linear, Slack, and Notion, catching drift before it reaches production.

If you've found misalignments in your codebase and want to be first in line: