Privacy Policy

This page is maintained by AmaQnA ("we", "us") to explain how we handle personal information across the AmaQnA website, product, JSON API, and embeddable widget. It is written in plain language and is not a substitute for legal advice.

1. Who we are

AmaQnA operates the website at amaqna.com and provides a voice-first Q&A service. Experts (hosts) publish a verified knowledge twin; audiences (askers) submit questions on the consumer feed, on a host's page, through a licensed embed, or via our JSON API. For privacy questions, contact [email protected].

2. Information we collect

From hosts

• Account: name, email, display name, slug, social links, bio, custom subdomain.
• Voice samples: audio you record or upload under an explicit voice consent and rights-holder warranty, used to create a private voice model.
• Corpus: text you paste, files you upload (PDF, DOCX, transcripts), and URLs you submit for ingestion; each item is chunked and embedded as a knowledge atom.
• Approvals & edits: your label choices (Personally approved, Source-grounded, Interpretive, Out of scope), edits, publish decisions, escalations, and mod actions — used to train your style guide.
• Payouts: for hosts that earn revenue share on Boosts, we collect the minimum identity and payout account information required by our payment processor.
• API keys: hashed credentials for publishers licensing your feed, with per-key usage logs.

From askers

• Questions: the text you submit to a host; automatically scrubbed for personally identifying information before it reaches the host.
• Optional email: only if you ask to be notified when your question is answered, subscribe to a host's digest, or purchase a Boost.
• Boost payment data: processed directly by our payment processor; we store only the transaction reference, amount, and refund status.
• Interaction data: hearts, follows, and daily streaks tied to an anonymous or account identifier.
• Technical data: IP address, timezone, and basic device info used for rate limits, abuse prevention, and product analytics.

3. How we use information

• Operate the service: create host pages, retrieve grounded context, draft answers with cross-family models, synthesize voice audio, and deliver notifications.
• Enforce trust: apply the four-state label to every answer, prepend the audible AI disclosure, and embed a C2PA-shaped provenance manifest in every synthesized clip.
• Gate publishing: block a host page from accepting questions until voice consent, a cloned voice, and a minimum approved knowledge base are in place.
• Prevent abuse: PII scrub, rate limiting, moderation queue, reports, and takedown workflows.
• Bill and pay out: process Boost payments, auto-refund questions that cannot be grounded, and pay host revenue share.
• Product analytics: aggregate metrics on question volume, plays, approvals, and cost telemetry.

4. Voice cloning, provenance, and consent

We do not create a voice model without an explicit on-record consent action by the host, plus a rights-holder warranty (you own or control the rights to the voice and are not bound by exclusive talent, label, union, or studio agreements that would prohibit this use) and an indemnity in our favour. Every synthesized clip opens with a short audible AI disclosure in the same cloned voice and ships with an ID3-embedded, C2PA-shaped provenance manifest plus a sidecar JSON. Hosts can revoke consent and delete their voice model at any time from the dashboard or by contacting [email protected]. Once deleted, we remove the model from our voice provider and stop generating audio for that host.

5. Subprocessors

We rely on carefully selected providers to run the service. Current categories and vendors include:

• Cloud hosting, database, storage, and auth: Lovable Cloud (Supabase infrastructure).
• AI drafting & embeddings: Lovable AI Gateway, which routes to at least two model families per answer (currently Google Gemini and OpenAI GPT).
• Voice synthesis: ElevenLabs.
• Web ingestion: Apify.
• Payments & payouts: Stripe (Boosts, subscriptions, host earnings).
• Transactional email: Resend.

We share only the data required for each provider to do its job, and we do not sell personal information.

6. Retention

We keep account data while your account is active. Questions and published answers are retained until the host removes them. Voice samples and voice models are retained until the host revokes consent or deletes them. Boost payment records are retained for the period required by tax and financial regulations. Logs and cost telemetry are kept for a limited period for security, billing, and debugging.

7. Your rights

Depending on where you live, you may have rights to access, correct, export, or delete your personal data, or to object to certain processing. Hosts can trigger a self-serve GDPR-style export from the dashboard. For any other request, email [email protected]. We will respond within the timeframes required by applicable law.

8. Children

AmaQnA is not directed to children under 13 (or the equivalent minimum age in your jurisdiction). We do not knowingly collect data from children. If you believe a child has provided us data, contact us and we will delete it.

9. International transfers

We may process data in countries other than your own. Where required, we use appropriate safeguards for international transfers.

10. Changes

We will update this page when our practices change and revise the "Last updated" date. Material changes will be highlighted on the site or by email where appropriate.

11. Contact

AmaQnA · [email protected]