LIVE during the demo

Find the blame.
Fix the outage.

BlameBot is the on-call agent that turns every Vercel deploy failure, Sentry error, and uptime alert into an explained incident, a paged owner, and (if you let it) an automatic rollback.

listening onVercelSentryUptimeRobotSlack
P1 / new

Production deploy failed: checkout/api

// AI report
TypeError: Cannot read properties of undefined (reading 'currency')
root cause: commit f3a1c removed currency fallback in cart adapter
blast radius: all checkouts since 14:32 UTC
recommend: rollback to previous deploy
rollbackhotfixautopilot
@sarah·billing pod · primary on-call
how it works

Four steps from alert to all-clear.

The whole loop runs in seconds. You stay in Slack. The bot does the digging.

step 01

Detect

HMAC-validated webhook receivers for Vercel, Sentry, and UptimeRobot. Deduped, normalized, never doubled-up.

step 02

Explain

Pulls the offending commit, files, and PR from GitHub. Surfaces the three most similar past incidents.

step 03

Triage

AI writes a tight report — root cause, blast radius, recommended action — and pages the right human.

step 04

Act

One reply in Slack runs the rollback via the Vercel API. Or let autopilot do it for you.

after every incident

Postmortems write themselves.

Once an incident is resolved, BlameBot writes a blameless postmortem from the timeline — summary, root cause, what went well, what went wrong, action items with owners and due dates.

Try the loop Click Load demo data on the dashboard.
postmortem.md
  • ## Summary
  • Checkout failed for 14m after a commit removed a fallback…
  • ## Root cause
  • currency adapter stopped guarding undefined cart…
  • ## Action items
  • [ ] add adapter unit tests · @sarah · 2d
  • [ ] gate cart adapter on canary · @priya · 5d