Subprocessors
Every outside company that processes personal data for the platform and its Startups, what each does, and how changes are announced.
A subprocessor is an outside company that processes personal data for the platform or a Startup. The table names every one, and each works under a data processing agreement.
The platform's subprocessors
| Subprocessor | Does | Data it processes |
|---|---|---|
| Cloudflare | Hosting, compute, storage, queues, network, and small-model inference | All platform data |
| ClickHouse | The analytics store that holds every Event | Event records |
| AWS | Model inference | Prompts and model responses |
| WorkOS | Sign-in, single sign-on, directory sync, and the vault for connected-account credentials | Names, email addresses, sign-in records and tokens |
| Stripe | Payments, connected accounts and identity verification | Payment, payout and identity details |
| GitHub | Repository sync, for a Startup that connects it | Code and repository metadata |
Each one's own security report is on file under Upstream review, and the Control opens a Task before any report expires.
Data upstreams
Some capabilities call outside data upstreams: enriching a Contact, reading the web, or generating media. Those that receive personal data run from a pinned allowlist, and each is named in the table before its first call.
Calls that carry no personal data, such as checking search rankings, are not subprocessing and are not listed.
A Startup's own list
Each Startup's list is the platform's list plus the Integrations it connects that receive personal data. It publishes with the Startup's paper and updates on every publish.
Changes
A new subprocessor appears in the table 30 days before it processes any personal data. Anyone can subscribe to changes. A Customer with a data processing agreement may object within those 30 days.
Next
- Your data: what happens to the data these companies process.
- Frameworks: which criteria are carved out to them.