Trust
Subprocessors
The third parties that process customer data on our behalf, and what each one does.
| Subprocessor | Purpose | Data processed |
|---|---|---|
| Google Cloud Platform | Infrastructure | Application data |
| Neon | PostgreSQL database | Application data |
| Cloudflare | Edge, WAF, object storage | Documents and traffic |
| Clerk | Authentication | User identity |
| Vercel AI Gateway | LLM request routing | Document text |
| Anthropic | LLM processing | Document text |
| OpenAI | LLM processing | Document text |
| Google Vertex AI | LLM processing and classification | Document text |
| Infisical | Secrets management | None |
| Sentry | Error monitoring | Application logs |
| Langfuse | LLM observability | Prompts (including document text) and outputs |
| Pipedream | Workflow automation | Google Drive sync events |
| Resend | Transactional email | Email addresses |
Processing and storage take place in the United States. Region details for individual subprocessors are available in our security pack on request.
We update this page when a subprocessor is added or replaced. To receive advance notice of changes by email, write to security@goodstream.com.
Last reviewed: September 10, 2026