Skip to content

Run SynapCores in Docker: docker run, Docker Compose and a first query

One container runs the whole engine: the database, the web UI, the REST API, the MCP server and the Ollama-compatible API, all on port 8080, with your data in a named volume. The image is multi-arch (linux/amd64 and linux/arm64), so the same commands work on Linux servers, Apple Silicon Macs and ARM cloud instances.

Tested on SynapCores 2.5.1-ceUpdated November 18, 2026

Install free, one command

curl -fsSL https://get.synapcores.com | SYNAPCORES_REF=docs-docker sh

Free Community Edition. Linux (x86_64, aarch64) and macOS on Apple Silicon; Windows via WSL2. Docker and other options

You need Docker (on macOS, Docker Desktop or colima), plus curl and jq for the first query. On Linux, prefix the commands with sudo if your user isn't in the docker group. Prefer a native install? Use the one-liner above and the quickstart.

Start the container

bash
docker run -d --name synapcores -p 8080:8080 \
  -e AIDB_ACCEPT_LICENSE=1 \
  -v synapcores-data:/var/lib/synapcores \
  synapcores/community:latest

The same image is mirrored on GitHub Container Registry as ghcr.io/synapcores/community:latest (same digests as Docker Hub), handy behind a proxy or when Docker Hub rate-limits pulls. Pin a release with a version tag such as synapcores/community:v2.5.1-ce.

Read the first-boot credentials

bash
docker logs synapcores 2>&1 | grep -A 7 FIRST-BOOT
Output
  FIRST-BOOT CREDENTIALS
===============================================================
  ADMIN LOGIN
  username: admin
  password: <24-character password>
  source: auto-generated
  Change it via the admin UI on first login.

The container prints this block once, on its first start. The full block also carries an API key and container notes:

Full first-boot block (your values differ)
===============================================================
  FIRST-BOOT CREDENTIALS
===============================================================
  ADMIN LOGIN
  username: admin
  password: <24-character password>
  source: auto-generated
  Change it via the admin UI on first login.

  OPENCLAW MEMORY (long-term memory plugin)
  OpenClaw Memory API key: <api key>
  → Use this to test the OpenClaw long-term-memory plugin immediately:
      openclaw plugins install @synapcores/openclaw-memory
    then set synapcores.apiKey to the key above.
  Capture this now — it will not be shown again.

  CONTAINER LIFECYCLE
  Read logs with docker logs synapcores; restart with docker restart synapcores.
  Keep the data volume and its private JWT/encryption keys across container replacement.
===============================================================

Log in and run a first query

This reads the password straight from the container log, logs in and defines a small sql helper for the REST API:

bash
PASS=$(docker logs synapcores 2>&1 | awk '/^  password: /{print $2; exit}')

TOKEN=$(curl -s localhost:8080/v1/auth/login -H 'Content-Type: application/json' \
  -d "{\"username\":\"admin\",\"password\":\"$PASS\"}" | jq -r .access_token)

sql() { curl -s localhost:8080/v1/query/execute -H "Authorization: Bearer $TOKEN" \
  -H 'Content-Type: application/json' -d "$(jq -n --arg q "$1" '{sql: $q}')" | jq -c '.data.rows'; }

sql "SELECT 1 + 1 AS two"
Output
[[2]]

Now store three notes, embed them inside the engine and search them by meaning:

bash
sql "CREATE TABLE notes (id INT PRIMARY KEY, body TEXT, emb VECTOR(384))"
sql "INSERT INTO notes VALUES
  (1, 'The cat sleeps on the sofa',  EMBED('The cat sleeps on the sofa')),
  (2, 'Quarterly revenue grew 12%',  EMBED('Quarterly revenue grew 12%')),
  (3, 'How to bake sourdough bread', EMBED('How to bake sourdough bread'))"
sql "SELECT id, body, COSINE_SIMILARITY(emb, EMBED('a kitten napping on the couch')) AS score
     FROM notes ORDER BY score DESC LIMIT 1"
Output of the SELECT
[[1,"The cat sleeps on the sofa",0.704431414604187]]

The search text and the best match share no keywords; the ranking comes from the embeddings.

On its first start the container downloads its embedding model (44 MB) and a local chat model sized to the memory Docker gives it. They live in the volume, so later starts skip the download. Check readiness with curl -s localhost:8080/health | jq -c '.ai | {state, ready}'; it reports {"state":"ready","ready":true} when both are loaded.

Open http://localhost:8080/ and sign in as admin to use the web UI.

Run it with Docker Compose

Save this as docker-compose.yml. Setting container_name keeps the docker logs synapcores commands above working unchanged:

docker-compose.yml
services:
  synapcores:
    image: synapcores/community:latest
    container_name: synapcores
    ports:
      - "8080:8080"
    environment:
      AIDB_ACCEPT_LICENSE: "1"
    volumes:
      - synapcores-data:/var/lib/synapcores
    restart: unless-stopped

volumes:
  synapcores-data:
bash
docker compose up -d
docker logs synapcores 2>&1 | grep -A 7 FIRST-BOOT

Then run the login and query commands from step 3. docker compose down stops the engine and keeps the volume; docker compose up -d brings it back with your data.

Keep your data

  • Everything the engine stores lives in the synapcores-data volume. Use docker restart synapcores to restart the engine in place.
  • Removing the container keeps the volume. Start a new container with the same -v synapcores-data:/var/lib/synapcores and it picks up the existing data: it skips the first-boot setup and reuses the downloaded models.
  • The volume also holds the engine's private JWT and encryption keys, so back it up together with your data.

Connect your tools

The container serves the same APIs as a native install. Point Claude Code, Cursor or VS Code at http://localhost:8080/v1/mcp, use the Python SDK, or call the Ollama-compatible API.

Next steps

Full reference

  • SQL reference: Every statement and function: EMBED, VECTOR, COSINE_SIMILARITY, GENERATE, Cypher, AutoML, agent memory.
  • API reference: The REST API. Your engine also serves its own OpenAPI schema at /v1/openapi.json.
  • System requirements: Supported platforms, memory and disk sizing.