One memory for every AI agent and every agent framework.

• Agents forget. Semurg remembers: facts and relationships written once, found again by the next agent, the next session and the next framework, each fact with the ids of the entities it connects.
• Your agent is the extractor: when it stores text it sends the entities and relations it found in the same call, so the graph grows with no schema to write first, and every answer comes back cited to the passage it came from.
• Every framework, one memory: LangChain and LangGraph, LangChain for Elixir, CrewAI, Microsoft Agent Framework, OpenAI Agents SDK, LlamaIndex, Mastra, Google ADK, Pydantic AI, the Vercel AI SDK and Jido have run end to end; Claude Code, Hermes Agent and OpenClaw connect with their own configuration (the integrations).
• Knowledge graph, vector search and graph RAG, built in: seven memory tools over MCP or plain HTTP (remember, document, recall, relate, forget, context and status), with vector search and graph RAG on the same copy of the data. Memory recall starts from the names you give it and walks their relationships.
• Bigger than RAM, on your own hardware: the engine's data lives on your disks, on-premises with no external cloud; built to grow past one node, licensed per node.
• One engine, written from scratch, for ten data domains: the memory is the same store that serves graph, vector, document, time-series, relational, columnar, key-value, object, streaming and search, in one on-disk format. Nine of the ten are measured in the open today against the engines teams use (graph, in memory and bigger than memory, vector, key-value, time-series, document, search, columnar, object and streaming, with vector + graph RAG: the benchmarks); relational is being measured on the current release and land there as they are measured, losses with the wins.

Try it in a minute with Docker on an Intel or AMD x86_64 Linux machine (WSL2 and Docker Desktop on Windows are being tested): start a node, remember a sentence, recall it.

docker run -d --name semurg -p 127.0.0.1:4000:4000 \
  --ulimit memlock=-1 --ulimit nofile=1048576:1048576 --security-opt seccomp=unconfined \
  -v semurg-data:/var/lib/semurg one.semurg.io/semurg
until docker logs semurg 2>&1 | grep -q 'Your API token'; do sleep 1; done
export SEMURG_API_TOKEN=$(docker exec semurg cat /var/lib/semurg/.api_token) NODE_URL=http://127.0.0.1:4000
mem() { curl -s -H "Authorization: Bearer $SEMURG_API_TOKEN" -H 'content-type: application/json' "$NODE_URL/v1/memory/$1" -d "$2"; }
mem remember '{"scope":"demo","content":"Ada Lovelace works at Acme."}'
mem recall '{"scope":"demo","subject":"Ada Lovelace"}'

Or install it on Ubuntu 24.04, as root: curl -fsSL https://one.semurg.io/install | sudo bash. A Mac version is coming. What it needs, and what it changes.