Docker Compose
Docker Compose is the fastest way to run Alquimia Platform locally. It starts the runtime together with the supporting services: Kafka, PostgreSQL, Redis, Qdrant, an S3-compatible object store (Alarik), Vault, and an OCI registry.
When to use this guide
Section titled “When to use this guide”- Local development and prototyping.
- Demonstrating the platform to stakeholders.
- Testing agent configurations before promoting them to a shared environment.
What runs
Section titled “What runs”| Service | Purpose |
|---|---|
| Runtime | Master (runtime, HTTP API) and worker (runtime-worker, agent execution) containers |
| Kafka | Event bus between master and worker roles |
| PostgreSQL | Persistent audit worklog and webhook subscriptions |
| Redis | Session state, caches, and locks |
| Qdrant | Vector store for knowledge bases |
| Alarik | S3-compatible blob storage for attachments |
| Vault | Secret storage and dynamic credentials |
| OCI registry | Storage for agent packages and Boltzmann Brains |
| Studio | Web UI for designing, testing, and managing agents (runtime 0.5.1+) |
Get started
Section titled “Get started”- Clone the
alquimia-runtimerepository. - Copy the example environment file and fill in the required secrets.
- Start the stack with Docker Compose.
- Verify the runtime health endpoint responds.
The Compose file runs the master (ALQUIMIA_RUNTIME_MODE=master) and a worker (ALQUIMIA_RUNTIME_MODE=worker) as separate containers that share the local registry directory, mounted read-only in the worker. For a single-process setup, run one container with ALQUIMIA_RUNTIME_MODE=all; this is convenient for local work but not recommended for production.
Environment setup
Section titled “Environment setup”At minimum you will need:
- An API token for authenticating requests.
- Database credentials for PostgreSQL.
- A Vault token for secret resolution (local Compose uses
VAULT_AUTH_METHOD=token). - Object-store credentials for the local S3-compatible store or S3.
- A signing key for CloudEvent integrity.
- An LLM provider API key if you want to run real inferences.
Store these values in the environment file; never commit secrets to source control.
Next steps
Section titled “Next steps”- Register your first agent
- Kubernetes & OpenShift — move to a production-grade deployment.
- Configuration — understand the runtime configuration model.