Privacy Policy

What this covers

This explains what Speakvora collects, why, who processes it and how long it is kept. It is written to match how the service works today.

What we collect

• Account: your email address and sign-in details (handled by Amazon Cognito), your API key records (we store a hash, not the key), and your settings.
• Usage: counts of characters, seconds of audio and agent turns, request times and error codes, used for billing, limits and support.
• Your input: text you send for speech, audio you send for transcription, agent questions and replies, and recordings you submit to make a custom voice.
• Billing: Stripe holds your card and billing details; we keep a customer reference and whether the account is paid.
• Site: standard server and CDN logs (IP address, browser, pages) for security and reliability. We do not run advertising trackers; pages load fonts from Google Fonts, which sees your IP address.

How long each thing is kept

• Catalog voice audio from /v1/speak is public by link; library clips are cached by the CDN for up to a year, and live-generated clips are kept so repeats are instant. The same text in the same voice returns the same file. Do not put personal data in catalog text.
• Custom-voice and agent audio is private to your account. It is stored once a sentence is used twice, kept 75 days after it was last used (each reuse restarts the 75 days), then deleted. Links are signed and last 5 minutes.
• Agent conversation memory is kept 24 hours.
• Audio you send for transcription is processed and not kept after the transcript is returned. A voice sample you submit is not kept after the voice is made.
• You can delete one voice, one recording, or all private audio from the portal at any time.
• Billing records are kept as long as tax and accounting law require.

Who processes it for us

We use these providers, each only to run the service: Amazon Web Services (hosting, storage, sign-in, email); DeepInfra (language model, transcription and custom-voice models; it states zero retention for the models we use); RunPod (GPU speech generation, which receives the sentence to speak); Stripe (payments). We do not sell your data and do not use your input to train models.

Your choices

You can access, correct or delete your data from the portal or by emailing support@eyeguide.ai; we respond within 30 days. Where privacy law gives you further rights (for example in the EU, UK or California) you can exercise them the same way. We use only the cookies and local storage needed to keep you signed in and remember your theme.

Your responsibilities

If you send other people's personal data (for example a customer's name in agent text) you are responsible for having the right to do so and for any notices the law requires. Custom voices must be your own voice or used with the speaker's consent.

Security and incidents

Data is encrypted in transit; stored data uses AWS encryption at rest. Private audio is reachable only through short-lived signed links. If we learn of a breach affecting your data we will tell you without undue delay.

Children

The service is not directed at children under 13 and accounts require you to be 18.

Changes

We will post changes here with a new date. Contact: support@eyeguide.ai.