Privacy Policy

Last updated: May 27, 2026

Your data stays yours

Flayr's three-layer privacy commitment. Last provider verification: 2026-04-11.

Flayr doesn't train on you.

We do not fine-tune, embed, cluster, or otherwise use your brand content, chat history, generated output, or uploaded assets to improve the product for other users. You're a customer, not a training dataset.

Our model providers don't train on you either.

Every AI call Flayr makes routes through Anthropic, Google Vertex, or OpenAI commercial APIs — all three contractually commit to not training on your inputs. Links to each provider’s own data-governance page are listed below.

You control retention.

Brand profiles, generations, and chat history can be deleted at any time from settings, with verified deletion across Convex, file storage, and any CDN cache. Your data is yours — including the right to remove it.

Provider citations

  • Anthropic Claude APIverified 2026-04-11
    “Anthropic will not use your Inputs or Outputs to train our models.”
    Read the full terms →
  • Google Vertex AIverified 2026-04-11
    “Google does not use your prompts or responses to train its models without your permission.”
    Read the full terms →
  • OpenAI APIverified 2026-04-11
    “Data submitted through our API will not be used to train or improve our models.”
    Read the full terms →

Deliberate carve-outs

Without these, the three-layer commitment would overstate what Flayr does. Here's exactly where the line sits:

  • Voice clone training uses your voice to build your clone.
    When you opt in to voice cloning, Flayr fine-tunes a per-user voice model on your recordings. That clone is stored only in your account, is never aggregated with anyone else's recordings, and is not used to train any shared model.
  • Generation outputs are not used to improve the product.
    Competitors sometimes feed generated content back into training via "feedback loops." Flayr does not. The outputs you keep belong to you; the outputs you delete are deleted.
  • We do log for debugging — but not for training.
    Error traces, API request metadata, and usage metrics are logged to Sentry and PostHog so we can fix bugs and size infrastructure. Those logs are never used to train any model, and they're scrubbed of brand content before long-term storage.

What we collect

Account information you provide (name, email, organization) through Clerk, content you generate or upload, usage metrics to improve the service, and billing information processed by Autumn.

How we use it

To provide the service, personalize AI generation, send product updates you've opted into, and comply with legal obligations. We do not sell your data.

Third-party services

We share data with: Clerk (auth + billing), Convex (database), PostHog (analytics), Vercel (hosting), Sentry (error monitoring), and AI providers (Anthropic, Google, Zernio) when you generate content.

Your rights

You can access, export, or delete your data at any time via your account settings. Contact privacy@flayr.social for requests.

Questions

Reach us at hello@flayr.social.