Voice is identity
A translation that strips your voice strips part of you. Cloning is not a gimmick — it's the point.
About us
We think the world should be able to talk to itself — in its own voice.
TranslationTone started with a simple observation: every translator on the market today flattens the speaker. Different language out, different person out. Phone calls between grandparents and grandchildren, doctors and patients, founders and customers — all of them turn into a generic robot the moment the language switches.
We believe that's the wrong default. The voice is the conversation. So we built a pipeline that runs translation and voice cloning across every channel you already use — live calls, voice messages, texts, and media — with the speaker's emotion preserved across the language boundary.
Today TranslationTone is an Android client, a web client, and a self-hosted FastAPI service that owns the GPU. Tomorrow it's iOS, then conferencing, then a developer API. The roadmap is long. The pricing is simple. The voice — yours — stays put.
What we believe
Four principles that drive every release.
Translate every channel
Calls, voice messages, texts, and media all live in one translated thread. No mode switching, no copy-paste, no "translate this" buttons.
Hand back the data
You own your voice profile and your transcripts. Training opt-in is genuinely optional — and off by default for sensitive audiences.
Quality is plural
There is no single right TTS engine. We expose the trade-off — speed, fidelity, naturalness — and let you pick per call.
Engineering
Built on the modern voice-AI stack.
Self-hosted translation, multiple voice-cloning engines (OpenVoice V2 and Chatterbox), faster-whisper for speech recognition, WebRTC for the live audio plane, FastAPI on the backend, SvelteKit on the web, native Kotlin on Android. Every piece chosen for one reason: it makes the conversation feel like a conversation.
- Kotlin / Android
- SvelteKit / TypeScript
- FastAPI / Python
- WebRTC + Socket.IO
- PostgreSQL
- CUDA / RTX-class GPU
- OpenVoice V2
- Chatterbox
- faster-whisper
- LibreTranslate
Want to be part of it?
Join the waitlist or get in touch.