A Real Graph Database for GBrain

What is GraphBrain?

GBrain stores its knowledge graph in Postgres — links are rows in a links table, traversal uses recursive CTEs. GraphBrain swaps the storage layer for Neo4j while keeping the exact same API. Every GBrain operation works unchanged, but now:

Connect GBrain to GraphBrain

Already have GBrain? Point it at GraphBrain's REST API in 2 steps.

curl -X POST https://graphbrain.belweave.ai/v1/brains \
  -H "Content-Type: application/json" \
  -d '{"name":"my-brain"}'Copy

# Set your API key as an env var
export GBRAIN_GRAPH_API_KEY=sk_your_key_here

# Point GBrain at your GraphBrain instance
gbrain init --engine graphbrain \
  --url https://graphbrain.belweave.ai/v1/brain_YOUR_BRAIN_IDCopy

For AI Agents

If you're an AI agent (Claude, Codex, Cursor, etc.) setting up GraphBrain for your user, copy-paste this:

Setup from scratch (no GBrain yet)

# 1. Install GBrain if not present
git clone https://github.com/garrytan/gbrain ~/gbrain && cd ~/gbrain && bun install

# 2. Create a GraphBrain brain
RESP=$(curl -s -X POST https://graphbrain.belweave.ai/v1/brains \
  -H "Content-Type: application/json" \
  -d '{"name":"my-knowledge-graph"}')
BID=$(echo "$RESP" | python3 -c "import sys,json; print(json.load(sys.stdin)['brain_id'])")
KEY=$(echo "$RESP" | python3 -c "import sys,json; print(json.load(sys.stdin)['api_key'])")

# 3. Initialize GBrain pointing at GraphBrain
cd ~/gbrain
GBRAIN_GRAPH_API_KEY=$KEY gbrain init --engine graphbrain \
  --url "https://graphbrain.belweave.ai/v1/$BID"

# 4. Start using it
echo "Brain ID: $BID"
echo "API Key: $KEY"Copy

Connect existing GBrain

# Already have GBrain set up with Postgres? Just point it at GraphBrain:
export GBRAIN_GRAPH_API_KEY=sk_your_key_here
gbrain init --engine graphbrain \
  --url https://graphbrain.belweave.ai/v1/brain_YOUR_BRAIN_ID

# Your existing scripts, MCP server, and cron jobs work unchanged.
# Graph operations (links, traversal) now use Neo4j.Copy

Run a smoke test

# Verify everything works
BASE="https://graphbrain.belweave.ai"
KEY="sk_your_key"
BID="brain_your_id"

# Health check
curl -s "$BASE/health"

# Create pages
curl -s -X PUT "$BASE/v1/$BID/pages/alice" \
  -H "Content-Type: application/json" -H "X-API-Key: $KEY" \
  -d '{"title":"Alice Chen","type":"person","content":"Engineer"}'

# Create links
curl -s -X POST "$BASE/v1/$BID/links" \
  -H "Content-Type: application/json" -H "X-API-Key: $KEY" \
  -d '{"from_slug":"alice","to_slug":"bob","link_type":"knows"}'

# Traverse
curl -s -X POST "$BASE/v1/$BID/traverse" \
  -H "Content-Type: application/json" -H "X-API-Key: $KEY" \
  -d '{"start_slug":"alice","depth":2}'

# Search
curl -s "$BASE/v1/$BID/search?q=engineer" -H "X-API-Key: $KEY"Copy

API Reference

All brain endpoints require X-API-Key: sk_... header. Cross-brain access returns 401.

Full API docs with request/response examples →

Roadmap

Deploy Your Own

One-Click Server Setup

curl -fsSL https://raw.githubusercontent.com/pkyanam/graphbrain/main/scripts/setup-server.sh | bashCopy

Works on any Ubuntu/Debian server with Docker. Installs Bun, Docker, Neo4j, GraphBrain, and creates a Cloudflare Tunnel in ~5 minutes.

Docker Compose (Local)

git clone https://github.com/pkyanam/graphbrain
cd graphbrain
docker compose up -d
# GraphBrain: http://localhost:3000
# Neo4j Browser: http://localhost:7474 (neo4j / graphbrain-dev)Copy

Full deployment guide (Railway, AuraDB, custom domains) →