AutoSRE is a multi-tenant SaaS platform that acts as your 24/7 autonomous on-call engineer. When a production incident fires (via PagerDuty, GitHub, or Slack), AutoSRE:

  1. Ingests the alert via webhooks or manual simulation
  2. Plans an investigation strategy using the Planner agent
  3. Analyzes logs, metrics, and deployment history in parallel
  4. Researches similar past incidents, CVEs, and internal runbooks
  5. Diagnoses the root cause with confidence scoring
  6. Acts — creates GitHub issues, Jira tickets, triggers rollbacks
  7. Communicates — posts structured Slack threads and email reports All of this happens in ~30 seconds, fully autonomously.

Built With

Share this project:

Updates