Subprocessors

Calafai uses a small set of third parties to deliver the service. When you submit a brief, run an engagement, or receive a deliverable, your data may pass through one or more of these subprocessors. We list them here in full so you can review the chain of custody before signing up.

Subprocessors fall into two groups: AI model providers, who run the language and image models that produce your deliverables, and operational subprocessors, the hosting, billing, email, and monitoring services that keep the platform running.

We will give at least 30 days' advance notice before adding a new third-party subprocessor that processes customer personal data, in line with our Data Processing Agreement (clause 4.2). Where we route to a new model from a provider already on this page, no new third party receives your data, and we update this page on the day of the change rather than 30 days ahead (clause 4.3). This lets us pick up newly released models from authorised providers without delay. The no-training commitment and the data-location representation must already cover the new model; if they do not, we treat the change as a new subprocessor and give the full 30-day notice. This page is the canonical record. Bookmark it, or subscribe to change notifications by writing to [email protected].

No training on your data. All AI model providers listed below are contractually prohibited from using your prompts, attachments, or completions to train, fine-tune, evaluate, or benchmark their models. We audit this annually. The LLM Provider DPA Audit is available on written request to [email protected].