AI-Powered · Free Early Access

Write incident post-mortems in seconds, not hours.

Run /postmortem in your Slack incident channel. BuildingAI pulls your Slack thread, PagerDuty timeline, and GitHub deploys — then drafts a structured RCA before the retro starts.

// Used by on-call teams who write post-mortems the same night, not the same sprint.

incident-postmortem.mdAI Draft
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
# Incident Post-Mortem

**Date**: 2024-09-17 03:42 UTC
**Severity**: P1  |  **Duration**: 47 min

## Summary
API gateway returned 503s at 03:14 UTC,
affecting checkout for ~18 min (94% error rate).

## Timeline
- 03:09 Deploy: api-gateway v2.4.1
- 03:14 PagerDuty P1: checkout 94% error
- 03:32 Rollback complete

## Root Cause Candidates
[hypothesis] Rate-limit config in
v2.4.1 too aggressive for prod traffic

## Action Items
- [ ] Add rollback test to CI pipeline
- [ ] Set pool alert threshold at 80%

*Draft — review before sharing externally.*
✓ Generated in 54s from 47 eventsSlack · PagerDuty · GitHub

Integrates seamlessly with your tech stack

💬Slack
🚨PagerDuty
⚙️GitHub
📋Jirasoon
📄Confluencesoon

Enterprise-Grade Security

Connecting Slack and GitHub is a big ask. Here is exactly what we do with your data.

We know the first question a security-conscious engineer asks is what happens to their incident data. The answer is simple: nothing persists.

📭

Stateless processing

Data fetched per-run, used to generate the draft, then discarded. Zero writes to any database.

👁️

Read-only permissions

OAuth scopes are read-only. The bot can never post, delete, or modify anything without your explicit action.

🔐

HMAC-verified requests

Every Slack event is verified against your signing secret. Unverified requests are rejected at the edge.

🏠

Your infrastructure

You deploy to your own Vercel account with your own tokens. We never have access to your credentials.

Audit the full source code on GitHub →// Open source · No hidden dependencies

How it works

Three steps from incident to signed-off RCA.

🔗
01

Connect

Link your Slack workspace, PagerDuty account, and GitHub repos in minutes. Read-only OAuth — we never write to your systems.

Setup takes ~15 min
⚡
02

Automate

Run /postmortem in any incident channel. The bot pulls the Slack thread, PagerDuty timeline, and GitHub deploys — then drafts the RCA in ~60 seconds.

60 second generation
📤
03

Export

Review the structured draft in Slack, edit inline, then export to Confluence, Notion, or markdown. Nothing auto-publishes without your sign-off.

You approve before publish
💬

Slack

Channel history + alerts

🚨

PagerDuty

Incident log (optional)

⚙️

GitHub

Commits & deploys

🤖

AI Draft

You review and edit

FAQ

Common questions

Does the bot read all my Slack messages?+

No. It reads only the channel where you run /postmortem, only for the time window you specify (default 6 hours). It never reads DMs or private channels.

Is my incident data stored?+

No. Each run is stateless. Slack messages, PagerDuty logs, and GitHub commits are fetched in the moment, passed to the AI, and discarded. Nothing is written to a database.

Do I need a paid Slack plan?+

No. The bot works on Slack's free plan using channels:history and chat:write scopes — both available on all tiers.

Is PagerDuty or GitHub required?+

Both are optional. The bot works from Slack history alone. Add PagerDuty and GitHub later by setting two environment variables — no changes to the Slack app needed.

Free Early Access · Limited Spots

Automate Your First
Post-Mortem Today.

Tell us about your team. We'll send step-by-step setup instructions and get you generating RCA drafts in under an hour.

Free during early access · No credit card · Setup in ~15 minutes

// No spam · We reply within 24h · Setup takes ~15 minutes