rebugr

Private beta · launching soon

Find the bug. Fix the code. Verify the result.

Rebugr traces an error to the exact line, writes a minimal patch, runs your tests in an isolated sandbox and opens a pull request you review.

waiting for an error
Error received
Cause traced
Fix proposed
Verified in sandbox
Sent for review
error
cause
fix
sandbox · npm test
Pull request is ready for review

One thread, from stack trace to merged fix.

The line on the left fills as you scroll. Each step is shown as it happens; nothing is assumed.

step 1Find

Paste an error or connect a repo. The stack trace is read, the files it touches are walked, and the failing line is pinned.

TypeError: cannot read properties
at formatTotal summary.tsx:88 ← here
at renderCart page.tsx:42
at CartPage page.tsx:17

step 2Understand

The likely root cause is named with the evidence for it, a confidence level, and what is still unknown.

cart.items is undefined on a new session
evidence: 3 traces, 1 commit, 0 counter-examples
side effects checked: formatTotal callers

step 3Fix

A minimal patch in a temporary workspace, with every change explained and nothing unrelated touched.

function formatTotal(cart) {
- return cart.items.reduce(sum, 0)
+ const items = cart?.items ?? []
+ return items.reduce(sum, 0)

step 4Verify

Your own test command runs in an isolated sandbox. What passed, failed or could not run is reported, then a pull request opens for you.

cart renders empty state
formatTotal handles no items
checkout totals match server
original failure no longer reproduces

Three places a bug shows up. One workspace to handle all of them.

Paste an error, get a tested fix.

Drop in a message, a stack trace or a log file. Rebugr explains it in plain language and proposes a targeted change.

  • Error and log analysis
  • Root cause with evidence
  • Patch with original and modified code
$ rebugr explain error.log
TypeError in cart/summary.tsx:88
items is undefined when the cart is new.
patch ready · 2 lines · tests 42/42

Failing check in, pull request out.

Rebugr reads the failed job, patches a temporary workspace, runs your tests and opens a pull request with the analysis attached.

  • GitHub repository integration
  • Isolated test execution
  • Never merges without your approval
ci: build #2231 failed
3 tests failed in cart.spec.ts
→ patch · sandbox run · 42 passed
pull request #412 opened for review

Live errors, linked to the commit that caused them.

Connect your monitoring. Incidents are grouped, matched to recent deploys and sent straight into the repair workflow.

  • Error monitoring integrations
  • Deploy and commit correlation
  • Incident history you can search
HTTP 500 rate +212% · /api/cart
linked to deploy v2.14 · commit a41c9f
→ sent to code repair
error rate back to 0.1% after merge

Same bug. Two ways to spend an afternoon.

Drag the handle. Left is the usual way; right is with Rebugr.

By hand~2 hours
    With Rebugr~4 minutes

      Automation that keeps you in charge.

      You approve every merge

      Rebugr opens pull requests. It never merges or deploys on its own.

      Your code runs sealed off

      Tests execute in an isolated environment with restricted repository permissions.

      Every result is shown

      Passed, failed or could not run, with the evidence. No green tick without a test behind it.

      Launching soon. Be first in.

      One email when the private beta opens. Nothing else.