Docker

Run Luna's server and web UI from a prebuilt image on any OS.

Run the image

terminal
docker run -d --name luna --restart unless-stopped \
  -p 127.0.0.1:8899:8899 -v luna_data:/app/data \
  ghcr.io/sehastrajit-s/luna:latest

Open http://localhost:8899 and pick a model in Settings. Images are published for linux/amd64 and linux/arm64 with every release, tagged :<version> and :latest.

⚠️
Keep the port on 127.0.0.1

Publishing it on all interfaces (-p 8899:8899) puts Luna on your whole network. To use it from other devices, share it privately over Tailscale instead.

Compose files

From a clone of the repo:

CommandRuns
docker compose -f deploy/compose.yml up -dLuna + Ollama on CPU
docker compose -f deploy/compose.yml -f deploy/compose.gpu.yml up -dLuna + Ollama on an NVIDIA GPU (needs the NVIDIA container toolkit)
docker compose -f deploy/compose.cloud.yml up -dLuna only, for a cloud model
docker compose -f deploy/compose.business.yml up -dBusiness mode: logins, rate limits, team channels. Opens the port on purpose; put it behind HTTPS.

The Compose files use the published image and fall back to building from your checkout. docker compose … up --build always builds locally. A .env in the repo root is optional and read if present.

Data and settings

Everything Luna writes lives in /app/data, so mount a volume there:

  • the database, memory, uploads, meetings and Library files;
  • settings saved from the Settings page, in /app/data/luna.env;
  • downloaded speech-to-text models, in /app/data/models;
  • the access token for remote devices, in /app/data/local_api_token.

The container runs as a non-root user, and nothing from your machine's .env is baked into the image.

Models

pull a model into the Compose Ollama
docker compose -f deploy/compose.yml exec ollama ollama pull qwen2.5:7b

Or choose a cloud provider in Settings → AI model and paste its key.

Upgrading

terminal
docker pull ghcr.io/sehastrajit-s/luna:latest
docker rm -f luna
docker run -d --name luna --restart unless-stopped -p 127.0.0.1:8899:8899 -v luna_data:/app/data ghcr.io/sehastrajit-s/luna:latest

# Compose:
docker compose -f deploy/compose.yml pull && docker compose -f deploy/compose.yml up -d

Build it yourself

terminal
docker build -t luna .
docker run -d -p 127.0.0.1:8899:8899 -v luna_data:/app/data luna

Every pull request builds the image and checks that it starts, answers its health check and serves the web UI.