Audio is processed for live translation
Microphone audio is sent through the real-time translation path so TolkTime can return translated voice and captions while the session is happening.
Trust
TolkTime is built for live translation, not for storing your conversations. This page explains what is processed live, what records we keep, who helps provide the service, and where automated translation has limits.
TolkTime processes media needed for a live session. We do not intentionally record or store conversation audio on TolkTime servers.
Microphone audio is sent through the real-time translation path so TolkTime can return translated voice and captions while the session is happening.
In Call mode, camera video and screen sharing use the browser call path. They are not sent to OpenAI as part of voice translation.
Live audio and translation output are processed by the service providers needed to run the session and remain subject to their technical and retention terms.
We keep the records needed to operate accounts, calculate translation usage, process purchases, show session history, maintain reliability, prevent abuse, and meet legal obligations.
For registered hosts, we store account details such as name and email, along with balance, purchase, receipt, and related billing records.
Session history can include the mode, selected languages, timestamps, duration, billed usage, and participant connection state. It does not include a stored conversation recording.
We may process browser and device type, operating system, IP address and approximate region, plus crash or error reports needed for security and reliability.
Call guests join through a browser link without creating an account. They should understand that their live speech will be processed for translation before joining.
Guests do not need an account, app install, subscription, or payment method. The host owns and pays for the Call session.
A guest chooses whether to allow microphone, camera, or screen access. Camera and screen sharing remain optional controls during the Call.
The host's history can include Call timing, selected languages, billed usage, and participant connection state, even though the guest does not create an account.
TolkTime uses specialist providers to run infrastructure, translate live audio, and process payments. Each receives the information needed for its role.
Cloudflare provides application infrastructure, storage, security services, and real-time Call transport.
OpenAI processes live microphone audio and produces translation output. TolkTime keeps OpenAI credentials server-side and does not expose them to browsers.
Paddle processes checkout, payment methods, receipts, and related billing information. TolkTime stores purchase and balance records but not full payment-card details.
TolkTime is designed to make live conversations easier to understand, but generated translation is not guaranteed to be complete or exact.
Names, accents, specialist terms, unclear speech, overlapping voices, background noise, connection quality, and the selected language pair can affect translation and captions.
Use the free balance to test your languages, selected voices, browser, microphone, speaker or headphones, network, and environment before a longer conversation.
Do not rely on automated translation alone for emergencies or consequential medical, legal, financial, safety, or compliance decisions. Use a qualified human interpreter where accuracy is critical.
Depending on your location, you may have rights to access, correct, delete, restrict, or receive a copy of personal data associated with your account.
Email support with the subject Data Request to ask for access, correction, deletion, restriction, or portability where applicable.
You may request account deletion, but records required for legal, fraud-prevention, security, tax, or accounting purposes may need to be retained.
This Trust page explains the product in practical language. The Privacy Policy contains the fuller description of collection, use, sharing, retention, transfers, cookies, and legal rights.
Start with 10 free minutes and test your languages, browser, microphone, playback, and environment in the mode you plan to use.