Install with Docker Compose

Before you begin

• Docker Engine 24+ (or Docker Desktop) with the docker compose plugin.
• The machine must be on the same network as your scanner (Uniden SDS200 or compatible), reachable over UDP 50536 and RTSP 554.

Install

1. Download toneout-compose.zip with the button above and unzip it somewhere permanent — recordings are stored next to it.
2. From the unzipped folder, run:

   docker compose up -d

3. Open http://localhost:6080 — or http://<server-ip>:6080 from another machine.
4. Finish in the browser: create your admin account, enter your scanner’s IP address, and paste your license key under Settings → License.

That’s it — the container restarts automatically with Docker after reboots.

NVIDIA GPU (optional)

If the host has an NVIDIA GPU and the nvidia-container-toolkit installed, run the CUDA variant instead for real-time large-v3 transcription (run docker compose down first if the CPU variant is already running):

docker compose up -d toneout-cuda

Configuration

Everything day-to-day — scanner IP, transcription model, alerts, users, SSL — is configured in the web UI under Settings. To pin a release version or change the ports, copy .env.example to .env and edit it.

Updating

docker compose pull
docker compose up -d

Your settings, database and recordings are kept — they live in Docker volumes and the recordings/ folder, not in the image.

Uninstalling

1. From the folder holding the compose file, stop and remove the container — keep the -v only if you also want to delete your settings and database (including your license key):

   docker compose down -v

2. Remove the downloaded image:

   docker image rm ghcr.io/jgilbs/toneout:latest

3. Delete the unzipped folder. Your recordings are in recordings/ inside it and are never deleted automatically — copy anything you want to keep first.

Running the CUDA variant? Use docker compose --profile cuda down in step 1. If you pinned a release with TONEOUT_VERSION in .env, remove that tag instead of latest in step 2.