Zapat

Getting Started

What is Zapat?

An overview of how Zapat automates the full dev lifecycle from issue to merged PR.

Updated March 3, 2026

Zapat is an AI-powered engineering automation platform. Label a GitHub issue or comment @zapat, and Zapat assigns an AI agent team to triage, implement, review, and open a pull request — automatically. The average issue costs $6 and completes in under 15 minutes.

How the pipeline works

  • You add a "zapat:implement" label to a GitHub issue — or comment @zapat on the issue.
  • Zapat receives the webhook within seconds and creates a job in the pipeline.
  • A triage agent reads the issue, classifies its type, and estimates complexity.
  • An implementation agent reads the relevant source files, writes the code, and opens a draft PR.
  • Review agents check the PR for security issues, code quality, and test coverage.
  • If all checks pass and auto-merge is enabled, Zapat merges the PR. Otherwise, it is ready for your human review.

What Zapat handles well

  • Bug fixes with clear reproduction steps
  • Feature implementations from well-described issues
  • Writing and extending test coverage
  • CI/CD failures — Zapat reads the error and applies a fix
  • PR code review with specific change requests
  • Research tasks and technical summaries
  • Rebasing branches that have fallen behind main

Tip

Zapat works entirely through GitHub — no separate CLI or local tooling required. Install the GitHub App, add a label or comment @zapat, and Zapat handles the rest.

The 10 trigger types

Zapat supports 10 distinct trigger types, each mapped to a specific workflow: triage, implement, review, rework, test, ci-fix, research, write-tests, rebase, and visual-verify. You activate them by applying the corresponding label to a GitHub issue or pull request — or by commenting @zapat on an issue or PR.

Connect your first repository

Was this article helpful?