Between translates spoken Cantonese, Mandarin, and English, and translates text in photos.
Microphone access is used only while a translation or transcript session is running. Leaving the app stops the session. Audio is sent over an encrypted connection to the speech services that produce the translation: Alibaba Cloud Model Studio (Qwen) for the Interpreter, Soniox for Transcript mode, and OpenAI through Vercel AI Gateway for beta Call mode. These providers process the audio to return text and speech. Between does not store conversations, and sessions request no provider-side conversation storage. Provider security and abuse-monitoring policies may still apply.
Camera access is used for the live camera view and for taking a photo in Camera mode. When you take a photo, Between sends it over an encrypted connection to OpenCode's inference service, which runs a vision model that reads and translates the text in the image. The photo is not stored on Between's servers. The captured image stays on the device only while the result is shown, and is discarded when you take another photo or leave the Camera tab.
The app keeps captions and photo translations in memory. Starting a new conversation, taking another photo, or closing the app clears them. Between does not include advertising or analytics SDKs, does not use data for tracking, and does not sell data.
Our server creates short-lived connection credentials and applies request limits. Service providers may process connection metadata needed to operate the service. No permanent service key is included in the app.
Software updates are downloaded from our server. Native changes are delivered through the App Store or TestFlight.