Engram

AI Memory Engine -- knowledge graph + semantic search + reasoning + learning in a single binary.

What is Engram?

Engram is a high-performance knowledge graph engine built as persistent memory for AI systems. It combines graph storage, semantic search, logical reasoning, and continuous learning into a single binary with a single .brain file.

• Single binary -- no runtime dependencies, no Docker, no cloud
• Single file -- one .brain file is your entire knowledge base. Copy = backup, move = migrate
• No external database -- everything is built in
• Hybrid search -- BM25 full-text + HNSW vector similarity + bitmap filtering
• Confidence lifecycle -- knowledge strengthens with confirmation, weakens with time, corrects on contradiction
• Inference engine -- forward/backward chaining, rule evaluation, transitive reasoning
• Ingest pipeline -- NER (GLiNER2 ONNX), entity resolution, conflict detection, PDF/HTML/table extraction
• Multi-agent debate -- 7 analysis modes with War Room live dashboard and 3-layer synthesis
• Chat system -- 47 tools across 8 clusters (analysis, investigation, reporting, temporal, assessment)
• Assessment engine -- Bayesian confidence with living assessments and evidence boards
• Temporal facts -- valid_from / valid_to on edges with automatic extraction
• Contradiction detection -- automatic conflict detection with resolution workflows
• Knowledge mesh -- peer-to-peer sync with ed25519 identity and trust scoring
• Built-in web UI -- Leptos WASM frontend with graph visualization, onboarding wizard, and SSE live updates
• Multiple APIs -- HTTP REST (230+ endpoints), MCP, gRPC, A2A, LLM tool-calling

Two Ways to Use Engram

Engram works in two complementary modes. Both use the same binary and the same .brain file.

Backend Engine (API / CLI)

Use engram as a headless knowledge graph engine. Store nodes, create edges, run queries, push inference rules, and ingest data -- all through CLI commands, HTTP REST, MCP, gRPC, A2A, or LLM tool-calling.

This is the integration path: embed engram into AI agents, automation pipelines, or existing applications. No browser needed.

engram create my.brain
engram store "Berlin" my.brain
engram relate "Berlin" "capital_of" "Germany" my.brain
engram serve my.brain

After starting the server, run through the onboarding wizard via API to configure your LLM, embedder, and NER providers. See the Configuration wiki for the API-based setup flow.

Interactive Playground (Web UI)

Start the server and 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.

Quick Start

1. Download

Download the latest binary from Releases.

Available for: Windows (x86_64), Linux (x86_64, aarch64), macOS (x86_64, aarch64).

2. Create and populate

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

3. Start the server

engram serve my.brain
# HTTP API: http://localhost:3030
# Web UI:   http://localhost:3030

4. Configure (recommended: Gemma 4)

We recommend Gemma 4 as the LLM (thinking mode, large context window). Run it locally with Ollama:

ollama pull gemma3:27b

Then configure engram via the web UI onboarding wizard, or via API:

curl -X POST http://localhost:3030/config \
  -H "Content-Type: application/json" \
  -d '{"llm_endpoint": "http://localhost:11434/v1/chat/completions", "llm_model": "gemma3:27b"}'

curl -X POST http://localhost:3030/config/wizard-complete

Any OpenAI-compatible LLM endpoint works (Ollama, vLLM, OpenAI, Azure, etc.).

CLI Reference

Documentation

Full documentation is available on the GitHub Wiki:

Use Cases

Built with Engram

License

Engram is free for personal use, research, education, and non-profit organizations.

Commercial use requires a paid license. Contact [email protected] for commercial licensing.

See LICENSE for full terms.