Quick start
Run llama-dash and llama-swap together with Docker Compose.
The fastest way to run llama-dash is with Docker Compose. Both setups bundle llama-dash with llama-swap, use ./config/config.yaml for the llama-swap config, mount ./models/ for model files, and expose llama-dash on http://localhost:3000.
Configure the environment
cp .env.example .envSet BETTER_AUTH_SECRET (generate with openssl rand -base64 33) and, if you'll store upstream provider credentials, CREDENTIAL_ENCRYPTION_KEY (32+ chars). The compose files load .env via env_file and won't start without it. See Configuration for the values to set first.
Create the llama-swap config
cp config/config.example.yaml config/config.yamlEdit config/config.yaml to declare your models. See the config file reference for the contract llama-dash expects.
Start the stack
Pick the compose file that matches your GPU vendor.
docker compose -f docker-compose.amd.yaml up -dRuns ghcr.io/mostlygeek/llama-swap:rocm, passes through /dev/kfd and /dev/dri, and also mounts /dev/dri into llama-dash so AMD GPU stats work in the dashboard.
docker compose -f docker-compose.nvidia.yaml up -dRuns ghcr.io/mostlygeek/llama-swap:cuda and requests gpus: all for the llama-swap service. This requires the NVIDIA Container Toolkit on the host.
Open the dashboard
Visit http://localhost:3000. On first launch you create the dashboard user, then you can load a model and send your first request.
llama-swap is not exposed on the host in the bundled compose setup — llama-dash
is the single public entrypoint on port 3000.
Next steps
Last updated on