Docker
Run Kowalski in Docker
Running Kowalski with Docker simplifies the setup, deployment, and management process significantly. Make sure you have Docker and Docker Compose installed, then continue to the next steps.
Running with Docker
Clone the Repo
git clone --recurse-submodules https://github.com/ABOCN/TelegramBotCopy a Docker Compose file
Without AI Features (Ollama)
mv examples/docker-compose.yml docker-compose.ymlWith AI Features (Ollama)
mv examples/docker-compose.ai.yml docker-compose.ymlSetup your .env
In order to successfully deploy Kowalski, you will need to have a valid .env. An example is provided in .env.example. You will need to edit both your .env file and enter matching values in webui/.env. Don't worry... it's mostly copy-paste!
You can learn more about the different options in the .env section.
If you intend to set up AI, the defaults for Docker are already included (just uncomment) and don't need to be changed. Further setup may be needed for GPUs. See the Ollama documentation for more.
Prepare the database
The WebUI and bot requires a Postgres database, which requires some setup. Again, don't worry as it's a pretty quick setup!
Migrate the schema to DB:
docker compose up postgres -d # Bring up just the database
bunx drizzle-kit pushRun the container
You're all done! You can run the container with:
docker compose up -dThe web interface will be available at localhost:3000. We suggest putting it a reverse proxy if you plan to use Kowalski for production use.