Real-time AI translation that keeps how you actually sound. For creators dubbing video, teams meeting across languages, and developers building speech products.
Whether you're translating a single video or shipping translation inside a product, start where you are.
Translate video, podcasts, and audio into 10 languages — keeping the original voice. $5 in credits free.
Try dubbing
Video meetings with a built-in interpreter. Invite anyone, speak your language, hear theirs.
Open a room
Live translation in Zoom, Meet, and Teams. Runs locally — no meeting bots, no extra invites.
Download previewMost translation tools strip a conversation down to words. Pinch keeps the voice, the timing, and the moments where people actually talk over each other — so a meeting in two languages still feels like a meeting, and a dubbed video still sounds like the person who recorded it.
We're an independent research lab building the speech models that make this possible, and shipping them in products you can use today.
Read our researchOne API for dubbing, real-time interpretation, and voice-preserving translation. Stream audio in, stream translated audio out. SDKs for Node, Python, and the browser.
// Dub a video — same voice, new language
const job = await pinch.dub.create({
source_url: "https://…/episode.mp4",
target_lang: "es",
preserve_voice: true,
});
// → job.output_url ready in secondsOccasional notes from the lab — research, releases, and demos. No spam.
Powered by Pinch API