Legal
Subprocessors
Last updated: May 28, 2026
To deliver Rufus we rely on a small set of third-party providers (“sub-processors”) who process personal data on our behalf. Each one operates under their own data-processing agreement with us, with appropriate security and transfer safeguards.
We'll update this page when we add, remove, or change a sub-processor. Material changes will be announced via our marketing channels at least 14 days before they take effect, giving you the chance to object.
Current sub-processors
Google LLC
United States · global- Purpose
- Sign-in (OAuth 2.0). We use Google as our identity provider — no Google Workspace data is read.
- Data processed
- Name, email, profile image, Google user ID.
Vercel Inc.
United States · global edge- Purpose
- Application hosting, CDN, serverless function execution.
- Data processed
- All data passes through Vercel infrastructure when serving the app, including authentication tokens and API payloads.
Neon Inc.
United States · region of your Neon project- Purpose
- Managed Postgres database — primary store for workspace content (knowledge base, proposals, RFPs, contracts, ledger).
- Data processed
- All user-generated content, account metadata, ledger.
Anthropic, PBC
United States- Purpose
- Large language model API (Claude) for generating proposals, answering RFPs, reviewing contracts, splitting knowledge-base imports.
- Data processed
- Only the inputs you explicitly send to a generation — e.g. the question text, the contract text, your knowledge base when it is needed as context for an RFP. Inputs and outputs are not used to train Anthropic's base models.
Stripe, Inc.
United States · global- Purpose
- Payments and subscription billing — checkout, recurring invoicing, credit-pack purchases.
- Data processed
- Name, email, billing address, card details (handled by Stripe — Rufus never sees the card number), purchase history.
International transfers
Several sub-processors are based in the United States. When personal data of EEA, UK or Swiss residents is transferred to them, we rely on the relevant transfer mechanism — typically the European Commission's Standard Contractual Clauses (and the UK addendum where applicable), supplemented by technical and organisational measures the sub-processor maintains under their DPA.
AI training opt-out
Anthropic, our AI provider, does not use customer API inputs or outputs to train its foundation models. Rufus inherits that posture: your workspace content is not used to train any AI model beyond the single inference call that delivered the output you asked for.
Questions
Contact privacy@exposql.com. Read the full Privacy Policy for how we use the data overall.