Open source, API based workflow orchestration tool.

This repo contains the API server, go client, and CLI.

Start a local server with the default config: houston api
Quickly run an end-to-end example workflow: houston demo
Or use the Docker container: docker run -p 8000:8000 datasparq/houston-redis demo
See the quickstart for a guide on how to create microservices and complete Houston missions using them: quickstart
You can download the binary from the releases page.
If you have go installed you can install with:
go install github.com/datasparq-ai/houston@latest
Houston is a simpler, faster, and cheaper alternative to tools like Airflow.
API based orchestration comes with 5 key advantages: 1. Code can run on serverless tools: lower cost, less maintenance, infinite scale 2. The server isn't under heavy load, so can handle hundreds of concurrent missions 3. Pub/Sub message delivery is guaranteed, improving reliability 4. Multiple workflows can share the same task runners, aiding collaboration 5. Task runners can run anywhere in any language, allowing for rapid development with no vendor lock-in
Please see the contributing guide.
Development of Houston is supported by Datasparq.
$ claude mcp add houston \
-- python -m otcore.mcp_server <graph>