Repo-aware deployment planning for AI-first teams.

Shipd reads a GitHub repository, scores hosting platforms against detected deployment signals, and produces a plan with blockers, tradeoffs, and step-by-step setup — before you touch infrastructure or change code.

→ Docs · Local setup

• Connects to GitHub with read-only OAuth access
• Scans deployment-relevant files: package.json, Dockerfile, CI workflows, env files, platform configs, infra folders
• Detects framework, runtime, topology, and deployment pattern
• Scores 11 platforms: Railway, Fly.io, Vercel, Render, Azure Container Apps, AWS App Runner, GCP Cloud Run, Heroku, DigitalOcean, Netlify, Docker + VPS
• Produces a saved deployment plan, platform comparison, and scan evidence trail
• Chat interface for follow-up questions, blockers, and tradeoffs

Shipd does not modify code, execute deployments, store secrets, or open pull requests.

• Node.js 20+
• PostgreSQL (local or remote)
• GitHub OAuth App
• OpenAI or Anthropic API key

git clone https://github.com/teckedd-code2save/shipd.git
cd shipd
npm install

Copy and fill in environment variables:

cp .env.example .env.local

Push the database schema and start the dev server:

npm run db:push
npm run dev

App runs at http://localhost:3000.

# Database
DATABASE_URL=postgresql://postgres:postgres@localhost:5432/shipd

# Auth
AUTH_SECRET=any-random-string
AUTH_GITHUB_ID=your-github-oauth-client-id
AUTH_GITHUB_SECRET=your-github-oauth-client-secret

# AI provider — pick one
AI_PROVIDER=openai
OPENAI_API_KEY=your-openai-api-key
OPENAI_MODEL=gpt-4.1-mini

# Anthropic (alternative)
# AI_PROVIDER=anthropic
# ANTHROPIC_API_KEY=your-anthropic-api-key
# ANTHROPIC_MODEL=claude-3-5-sonnet-latest

# Optional voice summaries (ElevenLabs TTS)
VOICE_ENABLED=false
ELEVENLABS_API_KEY=
ELEVENLABS_VOICE_ID=
ELEVENLABS_MODEL=eleven_turbo_v2_5

If DATABASE_URL is absent, auth falls back to JWT-only sessions with no data persistence.

npm run dev          # Start dev server (Turbopack)
npm run build        # Production build
npm run lint         # ESLint

npm run db:migrate   # Run Prisma migrations + regenerate client
npm run db:push      # Push schema to DB without a migration file
npm run db:generate  # Regenerate Prisma client only

Full documentation lives in /docs and is served by Docusaurus.

cd docs
npm install
npm start

GitHub files
    │
    ▼
scanRepositoryFiles()      ← deterministic parser (framework, runtime, signals)
    │
    ▼
classifyRepository()       ← repo class (web app, service, CLI tool, etc.)
    │
    ▼
matchArchetypes()          ← pattern matching (Next.js app, .NET solution, etc.)
    │
    ▼
scorePlatforms()           ← 11 platform rules → score + verdict + reasons
    │
    ▼
createPlanFromSnapshot()   ← deployment plan with blockers, warnings, next steps
    │
    ▼
PostgreSQL (cached)        ← served on repeat visits without re-scanning

AI extraction (extractRepoSignals) runs first and falls back to the deterministic pipeline if the LLM call fails.

See /docs for detailed architecture, scoring rules, and contribution guide.