AutoDoc

Pipeline-native AI documentation. PR summaries, ADRs, and onboarding guides — generated automatically, zero engineer input.

Every time a PR is opened or updated, AutoDoc reads the diff and generates:

• PR summaries — what changed and why, ready to share
• ADR drafts — architecture decisions captured while the context is fresh
• Onboarding updates — keeps new-engineer docs current automatically

Free for public repos. → autodoc.paprika-labs.app

---

Quick start

Add to .github/workflows/autodoc.yml:

name: AutoDoc

on:
  pull_request:
    types: [opened, synchronize, reopened]

jobs:
  autodoc:
    runs-on: ubuntu-latest
    permissions:
      pull-requests: write
      contents: write
    steps:
      - uses: paprika-org/autodoc@v0.1
        with:
          openai_api_key: ${{ secrets.OPENAI_API_KEY }}
          github_token: ${{ secrets.GITHUB_TOKEN }}

Add OPENAI_API_KEY to your repo secrets. That's it.

---

Example output

When a PR is opened, AutoDoc posts a comment like this:

📄 AutoDoc — Generated Documentation

What changed

• Added RateLimiter class in src/middleware/rate-limit.ts — sliding window, 100 req/min per IP
• Updated app.ts to apply rate limiting before auth middleware
• Added unit tests covering edge cases at the limit boundary

Why

Addresses production incident #147 — unauthenticated endpoints were vulnerable to scraping.

Testing notes

Verify rate limit headers (X-RateLimit-Remaining) are returned on every request.

---

🏛 Architecture Decision Record (draft)

ADR-DRAFT: Sliding window rate limiting at middleware layer

Status: Draft

Context: Need to protect unauthenticated API endpoints from abuse without per-route configuration.

Decision: Implement rate limiting as Express middleware using an in-memory sliding window counter, applied globally before auth.

Consequences: Simple to deploy; state is per-instance (won't work across pods without Redis). Acceptable for current single-instance deployment.

---

Configuration

Input	Default	Description
openai_api_key	—	Your OpenAI API key (required)
github_token	—	${{ secrets.GITHUB_TOKEN }} (required)
doc_types	pr_summary,adr	Which docs to generate: pr_summary, adr, onboarding
post_comment	true	Post docs as a PR comment
commit_docs	false	Also commit docs to docs/autodoc/ in the PR branch

---

Pricing

Free — public repos, self-hosted, bring your own OpenAI key. Full source on GitHub.

Pro — private repos, team analytics, custom output templates. Coming soon. Join the waitlist.

---

How it works

1. PR opens or updates → GitHub Action triggers
2. AutoDoc fetches the full diff via GitHub API
3. GPT-4o-mini analyzes the diff and generates structured docs
4. Docs posted as a PR comment (and optionally committed to the branch)
5. Your team reviews, edits, and approves before merging

---

Built by Paprika Labs

We build focused developer tools. autodoc.paprika-labs.app · paprika-lab.bsky.social