Getting Started with AutoKitteh: Build Your First Event-Driven Automation
A hands-on tutorial: run AutoKitteh with Docker, connect event sources, write a durable workflow in a few lines of code, and deploy it with full observability.
π Want to deploy AutoKitteh yourself?
Docker configs, system requirements, and installation guides β all on one page.
View AutoKitteh Tool Page βGetting started with AutoKitteh
In this tutorial you'll build your first event-driven automation with AutoKitteh. The only requirements are Docker and a few minutes of your time.
Step 1 β Run the server
Pull the official image and start the server. AutoKitteh listens on port 8080 and stores its state in a local volume, so your sessions survive container restarts:
docker run -d --name autokitteh -p 8080:8080 -v $(pwd)/ak_data:/data ghcr.io/autokitteh/autokitteh:latest
Step 2 β Connect an event source
Use the CLI to register a webhook trigger, or connect a built-in integration such as Slack or GitHub. AutoKitteh turns each incoming event into a session β a durable execution context that keeps its state even if the process restarts.
Step 3 β Write your workflow
Workflows are just code. A minimal example reacts to a webhook and calls an AI agent as a step:
def on_event(event):
result = ai_agent.invoke("Summarize: " + event.data)
return result
That's it β AutoKitteh handles retries, persistence, and observability for you. You can test everything locally with the VS Code extension before deploying.
Step 4 β Deploy and monitor
Deploy your workflow with a single CLI command, then watch execution traces in the web UI. If a step fails, AutoKitteh retries automatically with backoff β no more silent failures in cron logs.
π‘ Tip: Start with webhook triggers β they need no extra credentials and are perfect for learning durable execution before wiring up Slack or Kafka.
π Ready to run AutoKitteh?
Docker configs, system requirements, and installation guides β all on one page.
View AutoKitteh Tool Page β