How does the AI use your data?

Claude (Anthropic) writes your posts – in every workspace through Sayweek's own account, so you don't need a key of your own. OpenAI and Google Gemini are optional: we use them for transcribing voice notes, generating images and extra idea suggestions.

What the AI receives

• Only what the request needs: your brand description and rules, the channel, the topic and excerpts of your earlier posts.
• Your voice note, for the transcription – the audio itself is not stored.
• If you connected them: short, pre-filtered excerpts from your calendar, emails and CRM.
• Never passwords, tokens or payment data.

Mail, calendar, CRM: anonymised excerpts only

From mail, calendar and CRM sources only the subject, the date and a masked snippet are passed on. Contracts, prices, offers, health and legal matters are filtered out first, and from the CRM company and contact names, email addresses and amounts are removed. If a client name still remains in an idea, the idea is dropped. On Client names (open) you can list names that must never appear in a post.

No training on your data

Under their API terms, Anthropic, OpenAI and Google do not train their models on data received through their paid APIs. We don't train any general model with your data either, and we never pass it on to other customers. The style rules Sayweek learns from your posts and edits live only in your workspace.

You decide

Sayweek publishes nothing without your approval. AI can be wrong, so always check facts, names and numbers before publishing. For AI labels on posts, see AI transparency and compliance.