Sub-processors
- Last updated
- August 4, 2026
- Version
- 2026.08.3
- Applies to
- EU · France · Canada · United States
In plain language
- For core review and bring-your-own-key AI, Pyor has no server that sees your code. The app talks straight to GitHub from your machine, so your source, diffs, and token never reach a sub-processor. The one exception is the optional, opt-in “Pyor AI” feature: when you turn it on, the diff and changed files of a PR you run it on are sent through Pyor’s backend to our AI sub-processor, Azure OpenAI or Google Cloud Vertex AI (listed below), depending on the hosted model you select.
- The vendors below process only the limited personal data that running the website, the accounts/billing API, and analytics & diagnostics requires: billing (Polar), hosting (Vercel), usage analytics (PostHog, hosted in the EU; and Google Analytics on the marketing site only, consent-gated), and error diagnostics (Sentry). None of them ever receives your code, diffs, or token.
- This reflects our current stack. We keep it accurate, and we give notice before adding a sub-processor that processes personal data.
1. What is a sub-processor
A sub-processor is a third party that Pyor (“Pyor”, “we”) engages to process personal data on our behalf, under contract, to help us deliver the Pyor website and Services. Because Pyor is client-side for your code, for core review and bring-your-own-key AI, our backend receives only account and billing data, never your repositories, diffs, or reviews. The sole exception is the optional, opt-in Pyor AI feature, which sends the diff + changed files of a PR you run it on through our backend to Azure OpenAI or Google Cloud Vertex AI, depending on the hosted model you select (see the table below). Apart from that feature, sub-processors only ever handle the narrow set of personal data described in our Privacy Policy: what we need to run the site at pyor.review, operate accounts, and bill Teams.
We share data with a sub-processor only where it is necessary for one of those purposes, the vendor is bound by data-protection terms at least as protective as our own, and the transfer is covered by an appropriate safeguard. Card data is handled by Polar; we never see your full card number.
2. Current sub-processors
The table below lists the sub-processors that may process personal data for Pyor today: the confirmed, verified vendors behind billing, hosting, analytics, and error diagnostics. We engage no transactional-email provider yet; if we add one, it will appear here first under notice of changes. This is a living document.
| Sub-processor | Purpose | Data processed | Location / Transfer mechanism |
|---|---|---|---|
| Polar | Payments & billing: merchant of record | Billing contact and payment metadata. Polar handles all card data: we never receive card numbers and keep only a Polar customer reference. | United States / global: EU Standard Contractual Clauses (SCCs) |
| Vercel Inc. | Website + identity/billing API hosting and content delivery | Request logs (e.g. IP address, timestamps) for the marketing site and the accounts/billing API. Never your repositories, diffs, or token. | United States / global: SCCs |
| PostHog | Product & site usage analytics | Marketing site: aggregate, cookieless, consent-gated measurement with no personal profile. Signed-in app: product-usage events tied to your GitHub identity (user ID, login, email): never your code, diffs, or comment text. | European Union (EU region): Hosted in the EU; no transfer outside the EEA for this processor. |
| Google Analytics | Marketing-site usage analytics (consent-gated) | Marketing site only, and only after you opt in: aggregate page-usage measurement using first-party cookies (_ga / _ga_*) and a random client identifier: no advertising profile. Never your code, diffs, comment text, or signed-in app data; Google Analytics is never loaded in the app. | United States / global: EU Standard Contractual Clauses (SCCs) |
| Sentry | Error & crash diagnostics (stability and security) | Diagnostic data when something breaks: error messages, stack traces, and your GitHub identity (user ID, login, email) for triage. Tokens, secrets, and request/response bodies (which could contain code) are scrubbed before sending; healthy sessions are never recorded. | United States: SCCs |
| Microsoft Azure OpenAI | Optional first-party "Pyor AI" review assistance (opt-in) | ONLY when you turn on first-party Pyor AI and trigger a run: the diff and changed files of that pull request, sent through Pyor’s backend. Microsoft does not use it to train its or OpenAI’s models. Not used for BYO-key AI (which goes straight to your own provider) or for core review. | United States / global: EU Standard Contractual Clauses (SCCs) |
| Google Cloud Vertex AI | Optional first-party "Pyor AI" review assistance (opt-in) | ONLY when you select a hosted Gemini model and trigger a run: the diff and changed files of that pull request, sent through Pyor’s backend. Google does not use this data to train or fine-tune its AI/ML models without our permission or instruction. We have disabled Gemini’s default in-memory cache for our Vertex AI project; Google may still process prompts for applicable abuse monitoring. Not used for BYO-key AI, which goes straight to your own provider, or for core review. | Global Vertex AI endpoint (no fixed processing-region guarantee): SCCs where required. |
We do not engage sub-processors for advertising or cross-site tracking, and we set no ad cookies. Names shown as «…provider» are placeholders we will replace with the verified vendor: kept visible so an unfinished entry is obvious.
3. Notice of changes
We maintain this page as the authoritative list and keep it current. Before we add a new sub-processor that processes personal data, or materially change the role of an existing one, we will update this page and give reasonable advance notice.
- Teams customers can subscribe to changes by emailing support@pyor.review with the subject “Subscribe: sub-processor updates”. We’ll notify subscribers before a new sub-processor starts processing personal data.
- You may object to a new sub-processor on reasonable data-protection grounds. Write to support@pyor.review and we’ll work in good faith to address it; if we can’t, you may have remedies under your Data Processing Agreement.
4. International transfers
Some sub-processors process personal data outside your country, including in the United States. Where we transfer personal data out of the EEA, UK, Switzerland, or Canada, we rely on appropriate safeguards: chiefly the European Commission’s Standard Contractual Clauses (and the UK Addendum), together with a transfer-impact assessment and additional measures where needed. Business customers can sign our Data Processing Agreement, which incorporates the SCCs and lists the sub-processors authorised under it.
5. Contact
Questions about our sub-processors, or to subscribe to or object to changes: support@pyor.review. For the underlying processing, see our Privacy Policy.