← Spec Hub | Category: CRM & SaaS
Developer documentation for integrating Fireflies.ai ➔ Notion API Webhook Pipeline Specification webhooks into backend applications. Secure your endpoint with HMAC signature checking and payload validation.
Supported Events: contact.creation deal.stage_change company.updated ticket.created
Security validation specs for Fireflies.ai ➔ Notion API Webhook Pipeline Specification endpoints:
Authorization: Bearer sec_live_9784a80ba138ef02
const crypto = require('crypto');
const expected = crypto.createHmac('sha256', secret).update(rawBody).digest('hex');
const isValid = crypto.timingSafeEqual(Buffer.from(expected), Buffer.from(signature));
import hmac, hashlib
expected = hmac.new(secret.encode(), raw_body, hashlib.sha256).hexdigest()
is_valid = hmac.compare_digest(expected, incoming_sig)
{
"id": "evt_fireflies-to-notion.html_45396",
"event": "contact.creation",
"created_at": 1789764048,
"data": {
"resource_id": "res_251",
"status": "active"
}
}
curl -X POST https://your-domain.com/webhook/fireflies-to-notion.html \
-H "Content-Type: application/json" \
-H "Authorization: Bearer sec_live_9784a80ba138ef02" \
-d '{"id": "evt_fireflies-to-notion.html_45396", "event": "contact.creation", "created_at": 1789764048, "data": {"resource_id": "res_251", "status": "active"}}'
| Status Code | Status Meaning | Integration Action |
|---|---|---|
200 OK | Success | Event payload validated and queued. |
400 Bad Request | Malformed Payload | Check payload structure and JSON formatting. |
401 Unauthorized | Signature Mismatch | Verify secret key and digest computation. |
500 Error | Server Error | Triggers provider automated retry backoff. |
Always process incoming events asynchronously using a message queue. Never perform long-running DB queries or external API calls synchronously within the HTTP response handler.
Your framework might be auto-parsing JSON. HMAC generation requires the exact raw byte buffer of the request body before JSON middleware alters keys or spacing.
A 429 status indicates your endpoint is being rate limited. Implement dynamic throttling or increase consumer worker threads in your queue worker.
Store processed event IDs in an atomic storage like Redis with a set expiration. If an incoming ID exists, acknowledge with 200 OK immediately without re-triggering logic.