EngIntel

AI engineering intelligence platform: index GitHub repositories, chat over code context, and explore architecture, deployments, and observability from one workspace.

Stack

Layer	Tech
Frontend	Next.js 16, React 19
Backend	Express, TypeScript
AI	OpenAI (indexing, chat, summaries)
Telemetry	OpenTelemetry → Grafana Cloud (optional)

Monorepo workspaces: frontend/, backend/, packages/telemetry/.

Features

• Copilot — RAG chat over indexed repos
• Knowledge — semantic search and file index
• Architecture — system topology, service dependency map, Mermaid module graph
• Deployments — readiness checks and hosting hints
• Observability — OpenTelemetry metrics, latency, token/cost usage, traces, vector search

Prerequisites

• Node.js 20+
• OPENAI_API_KEY (required for indexing and chat)
• GITHUB_TOKEN (recommended for private repos and higher API limits)

Local setup

git clone <your-repo-url>
cd p1

cp backend/.env.example backend/.env
cp frontend/.env.example frontend/.env
# Add OPENAI_API_KEY, GITHUB_TOKEN, and BACKEND_URL=http://localhost:8000

npm install
npm run dev

• App: http://localhost:3000
• API: http://localhost:8000

Scripts

Command	Description
npm run dev	Frontend + backend (concurrent)
npm run build	Build telemetry package, backend, frontend
npm run start	Production start (both services)

Production deploy

Frontend (Vercel)

1. Import the repository root (not frontend/ alone) so workspace package @engintel/telemetry resolves.
2. Leave Root Directory empty (or .). vercel.json sets install/build for the monorepo.
3. Set env from frontend/.env.example, especially BACKEND_URL (your API URL).
4. GitHub OAuth: GitHub App callback https://repograph.shazeb.site/api/auth/callback/github. Vercel env: AUTH_URL=https://repograph.shazeb.site, AUTH_SECRET, AUTH_GITHUB_ID, AUTH_GITHUB_SECRET (or NEXTAUTH_* / GITHUB_* aliases).
5. Do not override install/build with a second npm ci in the build step — that can omit devDependencies and break PostCSS.

Backend (Railway / Render)

1. Deploy from repo root (see render.yaml) or backend/ as a Node service.
2. Set env from backend/.env.example, especially:
   • DATABASE_URL — Neon PostgreSQL pooled connection string (*.neon.tech). Remove any old Supabase URL from Render → Environment.
   • OPENAI_API_KEY, ANTHROPIC_API_KEY, GITHUB_TOKEN, CORS_ORIGINS / FRONTEND_URL
3. Start: npm start → node dist/index.js
   On Render use render:backend:build (includes prisma db push) and render:backend:start.

Indexed repositories, embeddings, and index jobs are stored in Neon PostgreSQL via Prisma (backend/prisma/schema.prisma). Set DATABASE_URL locally in backend/.env and in Render/Railway for production. Legacy JSON under backend/data/repos/ can be imported with npm run db:migrate-json --prefix backend.

Grafana Cloud (optional)

Export metrics, traces, and logs from the backend via OTLP:

OTEL_ENABLED=true
OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf
OTEL_EXPORTER_OTLP_ENDPOINT=https://otlp-gateway-prod-ap-south-1.grafana.net/otlp
OTEL_EXPORTER_OTLP_HEADERS=Authorization=Basic YOUR_BASE64_CREDENTIALS
OTEL_SERVICE_NAME=repograph-backend
OTEL_DEPLOYMENT_ENVIRONMENT=production
OTEL_RESOURCE_ATTRIBUTES=service.namespace=repograph
OTEL_TRACES_EXPORTER=otlp
OTEL_LOGS_EXPORTER=otlp
OTEL_METRICS_EXPORTER=none

On Render, also set NODE_OPTIONS=--require @opentelemetry/auto-instrumentations-node/register or use the backend start script (already includes -r).

Check: GET /health or GET /api/telemetry/status.

Health & smoke test

• GET /health — API up
• Index a public repo (e.g. vercel/next.js) in the UI to verify GitHub + OpenAI

mohammed.s.video.1.mp4

Troubleshooting

Issue	Fix
Backend offline in UI	Set BACKEND_URL in frontend env; check CORS_ORIGINS on API
GitHub rate limit	Add GITHUB_TOKEN to backend/.env, restart API, re-index
Empty architecture	Wait for indexing to finish, then open Architecture tab

License

Private — all rights reserved unless otherwise noted in the repository.

RepoGraph