Legal
Sub-processors
Version 1.0 - effective 5 August 2026
A sub-processor is another company we rely on to run Handbook that may touch customer data along the way. Here is the whole list - what each one does and where it runs. We keep this list short on purpose. If we plan to add or change one, admins get at least 30 days' notice by email first, as promised in the data processing agreement.
The list
| Service | What it does | Where |
|---|---|---|
| Anthropic | Generates answers from the question and the relevant document passages | United States - transient processing only, no training on your data |
| OpenAI | Turns documents into a searchable index (embeddings) | United States - transient processing only, no training on your data |
| Vercel | Runs the application | Compute pinned to London (lhr1) |
| Neon | The database - documents, knowledge base, transcripts and accounts | London (eu-west-2) |
| Stripe | Payments - card details go to Stripe directly and never touch us | Global payment infrastructure |
| Resend | Service emails, such as usage notices | United States, sending for handbookchat.com |
| Google Workspace | Our support inbox (support@handbookchat.com) | Google data centres (EU processing terms) |
The principles behind the list
Stored data stays in the UK and EU. The two AI providers see data transiently to do their one job each, under contracts that forbid keeping it or training on it. Identities for sign-in live in our own database rather than a hosted identity service, which is why you won't find one on the list. The reasoning is laid out in our privacy notice.