Homer
Checks your coding agent's evidence before it's allowed to say a task is done
Homer plugs into Claude Code, Cursor, or Codex and intercepts the moment the agent claims a task is finished, reading the actual evidence gathered and returning a calibrated confidence score plus the specific unfinished work. In an in-house comparison against the same agent without it, Homer produced 14% more correct final completions and a 44% reduction in premature or hallucinated ones, though without published methodology. The Free plan needs no credit card and gives 35% of the $20 plan's weekly usage; paid plans run $20 to $200 a month.
What holds up
- +Free plan needs no credit card and resets weekly with no setup beyond three minutes
- +Reported 44% reduction in premature or hallucinated completions in their own test
- +Works across three agents — Claude Code, Cursor, and Codex — with the same install flow
Mind the limits
- −14%/44% figures come from an in-house comparison with no published methodology
- −Paid usage doesn't roll over month to month, so unused allowance is lost
Picked here? Add the badge
Put it on your site. It links back to this page. Free for every listed product.
<a href="https://stillworks.dev/products/p/homer/"><img src="https://stillworks.dev/badge/homer.svg" alt="Picked by StillWorks" width="230" height="54"></a>