Run GrowthBook with Docker

The fastest way to run GrowthBook locally is with Docker Compose. The Compose stack below starts GrowthBook and a MongoDB instance together — no manual database wiring required.

🚀 Run GrowthBook with Docker first

Save this as compose.yaml, then start the stack:

# compose.yaml
services:
  mongo:
    image: mongo:latest
    environment:
      - MONGO_INITDB_ROOT_USERNAME=root
      - MONGO_INITDB_ROOT_PASSWORD=password
    volumes: ["mongodata:/data/db"]
  growthbook:
    image: growthbook/growthbook:latest
    ports: ["3000:3000", "3100:3100"]
    depends_on: [mongo]
    environment:
      - MONGO_URL=mongodb://root:password@mongo:27017/
    volumes: ["uploads:/usr/local/src/app/packages/back-end/uploads"]
volumes:
  uploads:
  mongodata:

Then run docker compose up -d and open http://localhost:3000.

No Docker? Use GrowthBook Cloud (free tier).

Step-by-step walkthrough

Save the compose.yaml shown in the box above to your project root. The file defines two services — mongo and growthbook — and two named volumes: one for MongoDB’s data directory, one for file uploads.
Start GrowthBook in the background by running:

docker compose up -d

The -d flag (detached mode) returns control to your terminal immediately while both containers start in the background.

Open your browser at http://localhost:3000 — this is the GrowthBook web UI where you create feature flags, run experiments, and manage your account.
Your SDK will talk to http://localhost:3100 — this is the GrowthBook SDK API endpoint. Your application fetches feature flag payloads from this URL, not from port 3000.
Stop and inspect the stack when you are done:

# Stop all containers (data is preserved in the named volumes)
docker compose down

# Tail live logs for the growthbook container only
docker compose logs -f growthbook

Full compose.yaml reference

services:
  mongo:
    image: mongo:latest
    environment:
      - MONGO_INITDB_ROOT_USERNAME=root
      - MONGO_INITDB_ROOT_PASSWORD=password
    volumes: ["mongodata:/data/db"]
  growthbook:
    image: growthbook/growthbook:latest
    ports: ["3000:3000", "3100:3100"]
    depends_on: [mongo]
    environment:
      - MONGO_URL=mongodb://root:password@mongo:27017/
    volumes: ["uploads:/usr/local/src/app/packages/back-end/uploads"]
volumes:
  uploads:
  mongodata:

Port reference

Port	Service	Purpose
3000	GrowthBook web app	Admin UI — create flags, experiments, metrics
3100	GrowthBook SDK API	Your application fetches feature flags from here
27017	MongoDB (internal)	Not exposed to the host; only used by the GrowthBook container

About MongoDB

GrowthBook stores all of its own configuration in MongoDB — feature flag definitions, experiment configs, metrics, and team settings. MongoDB does not hold your analytics event data or your user data warehouse. Your raw events stay wherever they already live (Snowflake, BigQuery, Redshift, etc.).

# Start GrowthBook and MongoDB in the background
docker compose up -d

# Stop all containers (data is preserved in the named volumes)
docker compose down

# Tail live logs for the growthbook container
docker compose logs -f growthbook

Trade-offs

Option	Benefit	Cost
Self-host with Docker Compose	Full control, no subscription cost	You handle ops yourself (backup, upgrades, uptime)
GrowthBook Cloud	No infrastructure to manage	Less control over your data and possible cost

Common mistakes

Running GrowthBook without persisting the MongoDB volume, so data is lost on container restart
Not setting critical environment variables (e.g. MONGODB_URI, JWT_SECRET) securely for production
Forgetting to back up the database before upgrading versions

💡 Real-world example

Small startups commonly start with self-hosted Docker Compose to save cost, then move to managed hosting once the team grows.