Privacy Policy
Last updated: 2026-11-16 · HookBridge processes inbound webhook events strictly on behalf of the authenticated workspace owner for inspection via the Dashboard and Model Context Protocol (MCP).
Privacy & Data Handling Summary
| Logging Policy | Zero-PII logging; request payloads are excluded from application logs |
|---|---|
| Ingest Processing | Bearer tokens, cookies, API keys, and secret prefixes are scrubbed before storage and re-checked on MCP reads (Dual-Layer DLP) |
| Secret Storage | Webhook signing secrets and outbound credentials use AES-256-GCM envelope encryption |
| Retention Schedule | 1 day (Free), 30 days (Pro), 90 days (Team) with asynchronous expiration purging |
| Model Training | Webhook payloads and action responses are never used to train foundation models |
Data Minimization & Inline DLP
- Header Sanitization: Sensitive headers (
Authorization,Cookie, API keys) and secret token patterns are redacted during ingestion before storage and sanitized again before MCP responses. - Boundary Isolation: Third-party strings returned to AI assistants are wrapped in
<untrusted_webhook_payload>tags to prevent indirect prompt injection.
Retention, Encryption & Non-Training Commitment
- Automated Retention Purging: Events expire automatically according to your workspace plan (1 day on Free, 30 days on Pro, 90 days on Team) and are purged asynchronously.
- Envelope Encryption: Webhook signing secrets and outbound gateway credentials are encrypted at rest using AES-256-GCM envelope encryption.
- No Training: Webhook payloads are never used to train foundation models.