How to Deploy Row-Bot with Docker: Step-by-Step Tutorial

In this tutorial, you will deploy Row-Bot — the local-first AI assistant for personal AI sovereignty — using Docker, in under ten minutes. No cloud accounts, no telemetry, no vendor lock-in. Just you and your own assistant.

Step 1 — Prerequisites

You need a Linux machine (or Docker Desktop on macOS/Windows) with Docker installed, plus 2 CPU cores and 4GB RAM as a minimum. For smoother voice and vision workloads, 4 cores and 8GB RAM are recommended. Optionally install Ollama to run models fully locally.

Step 2 — Pull the Image

Row-Bot publishes an official image on GitHub Container Registry. Pull it with:

docker pull ghcr.io/siddsachar/row-bot:latest

Step 3 — Run It

Create a directory for your data and start the container, mounting a volume so your knowledge graph and settings survive restarts:

mkdir -p ~/row-bot/data
docker run -d --name row-bot \
  -p 8080:8080 \
  -v ~/row-bot/data:/data \
  --restart unless-stopped \
  ghcr.io/siddsachar/row-bot:latest

Step 4 — Connect a Local Model

To stay 100% local, point Row-Bot at Ollama:

ollama pull llama3.1:8b
# set OLLAMA_HOST and ROW_BOT_MODEL in your env, then restart the container

Prefer the cloud? Row-Bot supports opt-in cloud models per component — just flip the setting in its UI and your local defaults remain untouched.

Step 5 — What to Try First

• 🗣️ Ask a question with your voice and let it reply out loud
• 🧠 Add a few notes and watch your personal knowledge graph grow
• ⏰ Schedule a daily task and let it run while you sleep
• 🌐 Hand it a URL and let browser automation summarize the page

Troubleshooting

That is it — a sovereign AI assistant running on your own hardware. For full requirements, the official compose file, and more deployment options, check the tool page below.