Privacy
Last updated 13 September 2026
The short version
If you are attending an event, we do not know who you are. There is no account, no name, no email address and no sign-up. Your phone remembers which language you chose; we do not.
What is spoken from the stage is sent to automatic transcription, translation and speech services to produce interpretation. The audio we generate is deleted within 24 hours. We do not store transcripts of what was said.
If you run an event, we hold your email address, your organisation, and records of what your events used so we can bill you.
Who is responsible for what
The organisation running an event decides what is spoken there and who is in the room. Under UK and EU data protection law they are the controller for that speech, and InfiniSpeak is a processor acting on their instructions.
For our own customer records — accounts, billing, enquiries — we are the controller.
InfiniSpeak is operated by Bethel Ventures LLC. Contact us at hello@everytongue.net.
If you are attending an event
We process:
- Speech from the stage. Not from your phone. Your microphone is never used, and nothing you say is captured — with two exceptions you control.
- Your own speech, only if you press Talk. Some events let attendees ask questions or join a discussion from their own phone. There, and only while you hold the Talk button, your microphone is used and what you say is interpreted live, exactly as speech from the stage is. It is not recorded or kept. A first name, if you give one, is shown to the room while you speak and is not stored. If you never press Talk, your microphone is never touched.
- What your phone hears, only while you use Listen for me. On your own, with no event, you can have your phone listen to a speaker and interpret them into your earbuds. Only while it is on, your microphone is used and what it hears is interpreted live, exactly as speech from the stage is. It is not recorded or kept, and it switches itself off after ten minutes with nothing said.
- Which language you chose, and whether captions are on — held on your own device, not on our servers.
- A count of how many people are listening in each language, so the operator can see the event is working. A count, not a list.
We do not ask for or receive your name, email address, phone number or location. We do not use tracking cookies, advertising identifiers or analytics that follow you. The only thing stored in your browser is your language choice and caption size, which you can clear at any time by clearing site data.
If you run an event
We process:
- Your email address, used to sign you in and to reach you.
- Your organisation name and the events, sessions and languages you set up.
- Usage records — how many language-hours each session used — which is what we bill against. These name the session and the language, not any attendee.
- Any glossary terms and pronunciations you add, which may include the names of people at your event.
If you contact us through the site we keep your name, email, organisation and message so we can reply.
Who else sees it
Interpretation cannot be produced without sending speech to specialist services. These are the ones we use, what each receives, and why. We do not sell data to anyone, and none of these services is permitted to use your event’s speech to train their models under the terms we hold with them.
| Service | Receives | For |
|---|---|---|
| Deepgram | Audio from the stage | Turning speech into text |
| OpenAI | That text | Translating it |
| Microsoft Azure | The translated text | Producing a voice |
| Cloudflare | The generated audio files | Delivering them to phones |
| Supabase | Accounts, events, usage records | Our database, in London |
| Fly.io | Live session traffic | Running the service, in London |
| Vercel | Web requests | Serving this website |
| Stripe | Your payment details | Taking payment — we never see your card |
For fifteen African languages, we run the voice model ourselves on our own machines. Speech in those languages is not sent to a third party for synthesis at all.
Some of these services process data outside the United Kingdom and the European Economic Area, including in the United States. Where that happens we rely on the standard contractual protections those providers offer.
How long anything is kept
| Generated interpretation audio | Deleted within 24 hours, automatically |
| Transcripts of what was said | Not stored |
| Records of individual listeners | Not stored |
| Listener counts | Aggregate only, kept with the event |
| Usage records for billing | Kept while we are required to keep financial records |
| Your account and organisation | Until you ask us to delete it |
| Enquiries | Two years, then deleted |
The 24-hour sweep runs on a schedule rather than a timer held in memory, so it survives a restart. We would rather delete audio a few hours late than fail to delete it because a process was restarted.
Your rights
If you are in the UK or the EEA you have the right to ask what we hold about you, to have it corrected or deleted, to object to how we use it, and to receive a copy. Write to hello@everytongue.net and we will respond within one month.
If your request concerns what was said at an event, the organisation running it is the controller and we will pass your request to them.
You can complain to the Information Commissioner’s Office in the UK, or to your own supervisory authority in the EEA.
Children
InfiniSpeak is sold to organisations, not to individuals, and accounts are for adults. Children may of course be in a room where interpretation is running — that is ordinary at a conference or a service — and nothing about listening identifies them or asks anything of them.
Changes
If we change how any of this works we will change this page and update the date at the top. If a change materially affects a customer we will email them before it takes effect.