pixwel CLI. Both run the same containers — API, MongoDB, Redis, ElasticSearch, ElasticMQ, and the background workers.
Codespaces (one-click)
Everything is preconfigured in.devcontainer/, so a Codespace boots the whole stack with no manual setup.
1
Open the Codespace
Open the repo as a Codespace in VS Code desktop (avoid the in-browser version). It requests an 8-CPU / 16 GB machine.
2
Wait for provisioning
The dev container builds and runs
./install.sh automatically (the postCreateCommand). This mounts the Docker socket and sets up the workspace at /workspaces/platform.3
Start the stack
Run
cd director; npm start. A dashboard opens with links to each running service.Local Docker setup
Prerequisites
- Docker
- Node.js
- AWS CLI, configured with credentials
- Linux only: raise the ElasticSearch mmap limit —
sysctl -w vm.max_map_count=262144(persist it in/etc/sysctl.conf)
Install
1
Clone and link the CLI
2
Add the environment file
Obtain
development.env and place it inside docker/, then export your GitHub packages token:3
Run the installer
4
Bring the stack up
http://localhost:8080.Load data
The stack starts empty. To seed it from a database dump:What’s in the stack
pixwel up runs docker-compose.yml (the app image is ghcr.io/pixwel/platform). The dev override (docker-compose.override.yml) exposes service ports for local access:
Other compose files layer on top:
docker-compose.minio.yml (local S3-compatible storage), docker-compose.debug.yml, and docker-compose.override.yml. Dockerfiles live under docker/ (docker/app/Dockerfile, Dockerfile-dev, docker/Dockerfile-ui, docker/mongo/, docker/elasticmq/).
Common commands
Running the UI
The frontend runs outside Docker:See also
- Architecture — how the API, UI, workers, and services fit together.
- Services — the background workers and what each does.
- Environments — staging, production, and how config differs.