Skip to content

Install SynapCores and run your first query

Install the engine with one command, read the admin password it prints on first start, then log in and run a SQL query and a semantic search from your terminal. A dozen command lines take you from install to the first ranked result.

Tested on SynapCores 2.5.0-ceUpdated November 18, 2026

Install free, one command

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

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

You need Linux (x86_64 or aarch64) or macOS on Apple Silicon, plus curl and jq (jq ships with macOS 15 and later; on Ubuntu run sudo apt install jq). On Windows, run every step in a WSL2 (Ubuntu) terminal; Microsoft's guide sets it up with wsl --install.

Install

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

The installer detects your platform, verifies the download checksum and sets up the engine:

  • Linux: installs the synapcores systemd service and starts it.
  • macOS: installs a few media libraries with Homebrew, then prints the command that starts the engine under Done — start the gateway. Run that command.

It also offers to download a local chat model sized to your RAM, so AI features work without an external API key.

Read the first-boot credentials

On its first start the engine creates the admin user with a random password and prints it once. On Linux, the installer's closing message points to the service log (press Ctrl+C to stop following it):

bash (Linux)
sudo journalctl -u synapcores -f

On macOS it appears in the terminal where you started the engine. The block looks like this:

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: aidb_<…>
  → 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.

Save the password now. The block is printed once, on the first start only. The same password also opens the web UI at http://localhost:8080/.

Log in and run your first query

Exchange the password for a session token, define a small sql helper that posts a statement to the REST API, and run a query:

bash
export SYNAPCORES_PASSWORD='paste-the-password-from-the-banner'

TOKEN=$(curl -s localhost:8080/v1/auth/login -H 'Content-Type: application/json' \
  -d "{\"username\":\"admin\",\"password\":\"$SYNAPCORES_PASSWORD\"}" | 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]]

Create a table with a VECTOR(384) column, let the engine compute embeddings with EMBED(), and rank rows by meaning with COSINE_SIMILARITY():

bash
sql "CREATE TABLE notes (id INT PRIMARY KEY, body TEXT, embedding VECTOR(384))"
sql "INSERT INTO notes (id, body) VALUES (1, 'Reset a forgotten password from the sign-in page'), (2, 'Download invoices as CSV from the Billing tab'), (3, 'Invite teammates from Settings, then Members')"
sql "UPDATE notes SET embedding = EMBED(body)"
sql "SELECT body, COSINE_SIMILARITY(embedding, EMBED('I am locked out of my account')) AS score FROM notes ORDER BY score DESC LIMIT 2"
Output
[["Table notes created successfully"]]
[[3]]
[[3]]
[["Reset a forgotten password from the sign-in page",0.437496542930603],["Invite teammates from Settings, then Members",0.09416782855987549]]

The search text and the best match share no keywords. EMBED() runs a local embedding model inside the engine, so the vectors never leave your machine and no API key is involved.

To clean up, run sql "DROP TABLE notes".

What to try next

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.