Two thumb buttons
One button for your language, one for theirs. Hold to speak, release to send. You run both sides of the conversation from a single phone.
Psybeam
Live Translator
Psybeam is a real-time voice interpreter for travel: you hold the phone and run both sides of the conversation. Hold your button and speak — they hear it aloud in their language. Hold theirs while they answer — you hear it in yours. The stranger in front of you never touches your phone.
./psybeam --screenshots
tap a screen to enlarge · swipe to see more → click a screen to enlarge
./psybeam --in-the-wild
cat psybeam/FEATURES.md
One button for your language, one for theirs. Hold to speak, release to send. You run both sides of the conversation from a single phone.
Natural speech-to-speech that follows the speaker's own voice. No typing, no reading a script, no teaching anyone the app.
Your words, in their language, rotate to face the person opposite you. Their reply stays upright for you to read.
Cross a border and Psybeam suggests the local language from your location, so you start talking instead of digging through menus.
Start with free minutes, then buy a pack only when you need more. No recurring charge and no trial trap.
Audio is streamed to OpenAI only to translate it. OpenAI does not use it to train models, though it may briefly retain API audio for abuse monitoring. Your transcript never leaves your phone.
man psybeam
Launch Psybeam and it goes straight to talking, no account or sign-up required.
Press your button and talk; Psybeam says it aloud in their language, then they hold their button to reply.
You hear their answer in your own language, face to face, in seconds.
less psybeam/README
Hand your phone to anyone and just talk — they never touch the screen.
Hold your button and speak; Psybeam says it out loud in their language. Hold their button while they answer, and you hear them in yours. The translation always faces whoever it's for, so the person across from you understands instantly — no app for them, no typing, no learning the UI.
Real-time voice-to-voice, built for travel. Order food, ask directions, sort out a hostel, talk to a pharmacist or a taxi driver — face to face, in seconds.
— No account, no sign-up. It opens straight to talking; sign in with Apple later only if you want to sync.
— No subscription, no trial trap. Pay only for the minutes you actually talk. Start with free minutes, then buy a pack when you need more — never a recurring charge.
— Your conversation stays on your device.
— Two thumb buttons: your language and theirs. Hold to speak, release to send.
— Natural speech-to-speech that follows the speaker's own voice.
— 22 languages, spoken aloud both ways.
— Cross a border? Psybeam auto-suggests the local language from your location.
— One phone, one operator — the other person just listens, then speaks when it's their turn.
Travelers, markets, hotel and hostel desks, taxis, pharmacies, asking a local for help — any moment you and someone else don't share a language.
iPhone, iOS 18 and later. Open it, hold a button, and have your first cross-language conversation in seconds.
./psybeam --faq
Psybeam is free to download and starts you with 5 free minutes. After that you buy a one-time minute pack, starting at $4.99, only when you need more. There is no subscription and no recurring charge.
No. Psybeam opens straight to talking with no sign-up. You can sign in with Apple later to link your account, but it is never required to use the app.
No. You hold the phone and run both sides. The person you are talking to just listens, then speaks when it is their turn, with an on-screen prompt inviting them in their own language.
22 languages, spoken aloud in both directions: Arabic, Chinese, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Polish, Portuguese, Russian, Spanish, Swedish, Thai, Turkish and Vietnamese.
Your conversation transcript stays on your phone and is never uploaded anywhere. To translate your speech, the audio itself is streamed to OpenAI in real time; OpenAI does not use that audio to train its models, though like any API traffic it may be retained briefly for abuse monitoring. Psybeam's own servers never see or store your audio or transcript.
No. Live voice translation needs a connection, since the speech is translated in real time by a cloud translator.
Psybeam runs on iPhone with iOS 18 and later.
Minute packs are handled through the App Store, so refunds follow Apple's standard purchase support process. There is no subscription to cancel.
Free to download with free minutes to start, then pay per minute with one-time packs, no subscription and no recurring charge.
ls ~/apps