Webhooks
Instead of polling, configure webhooks to receive results automatically. Provide Further AI with your webhook URL(s), authentication method, credentials if needed, and preferred response format.
Setup
To enable webhooks, provide the following to your Further AI representative:
- Webhook URL(s) — your HTTPS endpoint(s)
- Authentication method —
none,oauth2, orapi_key - Credentials — if using OAuth2 or API key auth
- Response format —
urlorinline
Authentication Types
Response Format
The response format controls how step data is delivered in webhook payloads.
URL Mode Step Data
Each step contains a signed URL to download the data:
Inline Mode Step Data
Each step contains the full data embedded directly:
Success Webhook Payload
Sent via POST to your webhook URL(s) when a workflow execution completes:
The steps[].data structure depends on the response format — either a download URL or inline data as shown above.
Failure Webhook Payload
Inline Payload Too Large
When using inline mode, if the payload exceeds the configured size limit (default: 1 MB):
When this occurs, the steps array is empty. Contact Further AI to increase the limit or switch to url format.
Retry Behavior
If your webhook endpoint does not return a 2xx status code, Further AI retries delivery:
Retry policy by error type:
- Network / transient failures — retried with exponential backoff
- 401/403 (OAuth2) — automatic token refresh, then retry
- Other 4xx errors — not retried (permanent client error)
- 5xx errors — retried with exponential backoff
Multiple URLs: Webhooks are sent to all configured URLs simultaneously, allowing you to notify multiple systems from a single execution.
Best Practices
- Return 2xx quickly — acknowledge the webhook immediately and process data asynchronously
- Handle duplicates — in rare cases a webhook may be delivered more than once; use
workflow_execution_idto deduplicate - Use HTTPS — HTTP URLs are not supported
- Log payloads — for debugging and auditing