Installation
How to get a running data pod — hosted, Docker, CLI, or from source
Installation
A data pod is your Synap backend: API, database, storage, identity, and jobs. Pick one path below; you can migrate later (export/import or reconnect clients).
After install: Quickstart routes you to the app, agents, or SDK.
Hosted: Synap Cloud
Best when you want to use Synap immediately without operating Postgres, search, or storage.
- Open synap.live and sign up (or sign in).
- Complete onboarding — you receive a workspace and pod URL in the product flow.
- Download the desktop app from the same landing / app shell when offered (browser works too).
Identity, provisioning, and security boundaries are described at a high level in Synap Cloud. You do not run docker compose for this path.
Docker
Best for local evaluation or a single-server install using the official images / compose stack.
- Clone github.com/Synap-core/backend (or your fork).
- Follow
deploy/in that repo: compose files, env templates, and thesynapshell script for install / update / migrate when you run from source or image. - Deep checklists (production, TLS, backups) → Architecture → Deployment.
Typical local flow (abbreviated):
Exact service names, ports, and health checks change with releases — treat the repo deploy/ docs as canonical; this page stays the map.
CLI
Today there are two CLI-related entry points; we are centralizing them so one story covers pod ops and agent bridge.
| Surface | When to use it |
|---|---|
synap script (in the backend repo, e.g. deploy/../synap) | Install, update, migrate, and operate the data pod from Docker/source on a server you control. |
npx @synap/cli (npm) | Connect OpenClaw, laptops, and external agents to your pod (guided init, Hub-oriented flows). |
Rule of thumb: server lifecycle → repo synap; “connect my AI stack” → npx @synap/cli.
Agent connection details → AI agents & external tools.
From source
For core development on Synap itself (API, workers, schema):
Prerequisites
- Node.js 20+ and pnpm (version pinned in repo)
- Docker Desktop for local dependencies (Postgres, MinIO, Redis, Ory, …)
- Optional provider keys (OpenAI, Anthropic, …) for AI features
Configure environment
Start dependencies
Common services: PostgreSQL (+ Timescale / pgvector), MinIO, Redis, Kratos, Hydra — see repo docker compose.yml for current list and ports.
Database & app
Verify with the repo’s health endpoint (path/port in README).
Production
Use Production checklist and Deployment overview — not a duplicate of those guides here.
Troubleshooting (local Docker)
Port clashes: change PORT in .env or free the port with lsof -i :PORT.
Next
- Quickstart — app, agents, or SDK
- AI agents & external tools — OpenClaw & Hub Protocol
- Architecture → System concepts — how it fits together
