Dashjoin can be run in the following configuration
- Platform: the main Dashjoin Platform
- AIA: the AI backend
- MCP-Proxy: Allows installing MCP tools to be used by AIA
- Postgres: the database for both platform and AIA
First, clone the GitHub repository to get access to the environment:
git clone https://github.com/dashjoin/platform
cd platform/docker-compose
If you are running ollama (http://host.docker.internal:11434) locally, you are good to go. If you want to use an external AI service, please enter your API key in the .env file.
The files mcp-proxy-config.json and settings.yml contain the default setup for your AI models and MCP servers. Usually, there is no need to make changes here. Please refer to the documentation for instructions regarding configuring custom models.
To start, stop, and delete the setup, run these commands:
# start containers
docker compose up
# stop containers
docker compose stop
# delete containers
docker compose down
In your working directory, you will find a folder called "volumes". It contains the persistent volumes for the dashjoin, aia, and postgres containers. Note that the data is retained even after "docker compose down" is run.
You can run MCP servers installed in the MCP toolkit in docker desktop. To expose them to clients, run this command:
source .env
export MCP_GATEWAY_AUTH_TOKEN=${MCP_GATEWAY_AUTH_TOKEN}
docker mcp gateway run --transport=http