Quickstart
Use the FlowPulse ingestion API from n8n, Make, Zapier, or any HTTP client to report when automation runs succeed or fail.
How FlowPulse works
FlowPulse does not run your automations.
Your workflows send run events (success or failure).
FlowPulse stores those runs and provides visibility, alerts, and history.
Important: FlowPulse only knows about runs you send.
If your workflow does not send an event, FlowPulse cannot track it.
When to use FlowPulse
Use FlowPulse if you want to:
- Track runs across multiple tools.
- Detect failures and silent runs.
- Centralize monitoring.
Integration mental model
You typically add FlowPulse at the end of your workflow:
- Success path -> send success event.
- Failure path -> send failure event.
Example workflow
Example: detecting bad AI output
- LLM generates output.
- Validate output (schema / rules).
- If invalid -> send failure event.
- If valid -> send success event.
What to send
Recommended fields:
- Execution ID.
- Duration.
- Input/output references.
- Error message.
- Environment (
prod/dev).
What FlowPulse does not do
FlowPulse does NOT:
- Run your workflows.
- Retry failed executions.
- Evaluate AI outputs automatically.
FAQ
Does FlowPulse detect hallucinations?
No. FlowPulse does not automatically detect hallucinations or bad outputs.
You define what "bad" means in your workflow (for example, validation or guardrails).
Then you send that result to FlowPulse.
Your integration sends HTTP POST requests when a run finishes. FlowPulse stores one run per execution.
Before you start
-
Access to a FlowPulse workspace.
-
A workflow created in FlowPulse under Workflows. You need its
workflow_idorworkflow_slugfor API calls. -
An ingestion API key from Developer or workspace settings in the FlowPulse app. Keep it in server-side or secure automation config, not in public browser code.
API base URL
Ingestion uses the same origin as the FlowPulse dashboard.
| Origin | https://app.flowpulse.xyz |
| Ingestion API | https://app.flowpulse.xyz/api/v1 |
Append paths such as /ingest/run-success to the ingestion base.
Send your first event
Every request:
- Method:
POST - URL:
https://app.flowpulse.xyz/api/v1/ingest/run-successorhttps://app.flowpulse.xyz/api/v1/ingest/run-failure - Headers:
X-API-Key: <your key>orAuthorization: Bearer <your key>, andContent-Type: application/json - Body: JSON that identifies the FlowPulse workflow and includes timestamps (see API reference)
Semantics:
success= your workflow completed as expected.failure= anything you consider incorrect (error, timeout, invalid output).
Identify the workflow with one of:
workflow_id: stable ID from the FlowPulse UI orGET /api/v1/workflows. Best for production.workflow_slug: URL-friendly slug from the workflow page (for examplelead-enrichment). Fine for tests and internal tools.
Minimal success example (replace YOUR_API_KEY and YOUR_WORKFLOW_SLUG):
curl -X POST "https://app.flowpulse.xyz/api/v1/ingest/run-success" \
-H "X-API-Key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"workflow_slug": "YOUR_WORKFLOW_SLUG", "ended_at": "2026-04-06T12:00:00.000Z"}'
Next steps
- Authentication (header formats)
- API reference (full fields and errors)
- Code examples (curl, Node.js, Python)
- Integrations (n8n, Make, Zapier)