Workflows
Chain multiple extractions into an automated pipeline
A workflow connects multiple playbooks into a pipeline. Use workflows when you need data from several websites in sequence, or when the output of one extraction feeds the next.
What workflows can do
- Run extractions from multiple websites in sequence or in parallel
- Pass data from one step to the next (e.g., extract URLs from Site A, then visit each on Site B)
- Schedule the entire pipeline with a cron expression or webhook trigger
Creating a workflow
From the Workflows page, click New Workflow. The workflow builder lets you:
- Add playbook nodes — each node runs a saved playbook
- Connect nodes — define execution order by drawing connections between nodes
- Map data — pass output fields from one node as inputs to the next using
{{stepName.fieldName}}syntax
Trigger types
| Trigger | Description |
|---|---|
| Manual | Click Run in the UI |
| Scheduled | Cron expression — e.g., 0 6 * * * for daily at 6 AM |
| Webhook | POST to the workflow's webhook URL from any external system |
Example: Cross-site odds comparison
Comparing match odds across two betting sites:
- Node A — runs a playbook that extracts odds from Site A
- Node B — runs a playbook that extracts odds from Site B (parallel to A)
- Workflow collects both outputs for downstream processing
Configure the trigger as 0 9 * * * to run every morning at 9 AM automatically.
Workflow runs
A workflow run tracks execution across all nodes:
- Which nodes have completed, are running, or are pending
- The accumulated data from each step
- Any node-level errors
You can watch node progress individually — each node's live view is available while it's running.
Notes
- Playbooks must be saved and working before they can be added to a workflow node
- Workflows require at least one trigger node to be runnable on a schedule
- A workflow run fails if any required node fails; optional nodes can be configured to continue on error