Getting Started

Getting Started

Download

Download the latest binary from Releases.

Platform	File	GPU Acceleration
Windows x86_64	engram.exe	DirectML (NER/RE)
Linux x86_64	engram-linux-x86_64	CUDA (NER/RE)
Linux aarch64	engram-linux-aarch64	CUDA (NER/RE)
macOS x86_64	engram-macos-x86_64	CoreML (NER/RE)
macOS aarch64	engram-macos-aarch64	CoreML (NER/RE)

All builds include wgpu compute (DX12/Vulkan/Metal) for similarity search acceleration regardless of the NER GPU feature.

Create Your First Brain

engram create my.brain
engram store "PostgreSQL" my.brain
engram store "Redis" my.brain
engram relate "PostgreSQL" "caches_with" "Redis" my.brain
engram query "PostgreSQL" 2 my.brain

Start the Server

engram serve my.brain

This starts the HTTP API on http://localhost:3030 and serves the web UI at the same address.

Mode 1: Backend Engine (API / CLI)

Use engram as a headless knowledge graph. No browser needed.

# Store from CLI
engram store "Berlin" my.brain
engram relate "Berlin" "capital_of" "Germany" my.brain

# Or use the HTTP API
curl -X POST http://localhost:3030/store \
  -H "Content-Type: application/json" \
  -d '{"entity": "Munich", "node_type": "city"}'

After starting the server, configure your LLM and embedder via API calls. See Configuration for the headless setup flow.

Mode 2: Interactive Playground (Web UI)

Open http://localhost:3030 in your browser. Four sections:

• Knowledge -- graph explorer, search, documents, facts, chat
• Insights -- intelligence gaps, assessments, contradictions
• Debate -- multi-agent analysis with 7 modes and a live War Room
• System -- configuration, NER/RE settings, sources, domain taxonomy

On first launch with an empty brain, the onboarding wizard guides you through 11 setup steps.

MCP Server

Use engram as a tool inside Claude Code, Cursor, or Windsurf:

engram mcp my.brain

Add to .mcp.json:

{
  "mcpServers": {
    "engram": {
      "command": "engram",
      "args": ["mcp", "/path/to/my.brain"]
    }
  }
}

See MCP Server for the full tool reference.

Next Steps

1. Configure your LLM, embedder, and web search
2. Read the HTTP API reference
3. Try a Use Case