Face-to-face conversations
Two people, one phone: hold, speak, then pass the turn.
If the person you want to speak with is beside you, a telephone call adds a step you do not need. A shared-phone translator gives each person a clear turn: choose their language, hold their button while they speak, then release it to translate. Here is a simple routine for a short conversation at a counter, a reception desk, or around a table.
Choose languages together before starting
Ask which language the other person prefers instead of choosing from their appearance or nationality. Put that language on Person 2 and yours on Person 1. Show them which button belongs to them before the first question.
Both people should agree to the translation process. Origin’s face-to-face mode is for adults 18 and older. Its permission screen explains that recorded turns go to Deepgram for transcription and the text goes to OpenAI for translation. Ask again when you start a conversation with a different person.
Release means “translate this turn”
Hold your button, say one short thought, and release. Keeping it pressed through a long silence keeps recording; releasing tells the app that your turn is ready to translate. Try a single question such as: Which platform does the airport train leave from?
Wait while the translation plays. Then the other person holds their own button and replies. In this mode the microphone stops before playback, so the app is not continuously listening while its translated voice speaks. Nearby people can still interrupt a recording, so choose a quieter spot when possible.
Check the meaning before adding another question
Look at the original text as well as the translation. If the original is wrong, repeat a shorter phrase. If an answer seems unrelated, clarify the current question before moving on; adding another question can make the misunderstanding harder to trace.
Confirm important details in small pieces. For example: Platform six? Then: Does it stop at the airport? Translation and recognition can make mistakes, and neither a fluent voice nor a neat caption proves that every detail is correct.
Use text when the room or microphone gets in the way
Type is a useful alternative if microphone permission is unavailable or background noise keeps changing the transcript. Each person types in their selected language and reviews the translated text. Listen replays a completed translation without recording another turn.
If speech cannot play, read the captions and check your device’s available voices. Origin’s face-to-face mode uses standard device voices; it does not clone either person’s voice. Speech availability varies by browser, device and language. An internet connection is required.
Choose the mode that matches where the other person is
We build Origin at Avatar8 LLC. The shared-phone web mode described here is available now. It needs no account or telephone number. It is a turn-taking interface, not simultaneous overlapping interpretation. Origin does not support emergency calling.
As of November 15, 2026, the new native iPhone face-to-face interface is awaiting external TestFlight review. The public App Store listing below describes Origin’s existing translated phone-call functionality. Do not assume installing that public version already provides the new interface. Telephone calling remains a separate feature; weekly and monthly number rentals are not available.
Origin does not archive these face-to-face turns on its server. Text remains in the current screen until cleared or the screen closes; processing providers have their own retention policies. Review the privacy policy together before using the microphone.