About LangsAny

Translation that stays with you.

LangsAny exists because of one conviction: you shouldn't have to hand your words to a server to get them translated. So we moved the translator itself — open-source AI models, running in your browser, on your device.

Every mainstream translation site computes in the cloud: your text is uploaded, processed in a data center, and returned. That design forces three compromises — your content leaves your control, every request costs the provider money (so you get limits), and no connection means no translation.

LangsAny inverts it. A compact neural translation model is downloaded to your browser once, cached, and runs locally from then on. Private, free, unlimited, and offline-capable are not four features we added — they are four consequences of a single architectural decision about where the computation lives.

Private by architecture

Translation runs in your browser. Your text never reaches a server — not because we promise, but because no upload path exists.

Free & unlimited

Your device does the computing, so we have no per-request costs. No character caps, no quotas, no account.

Offline-capable

Models are cached on your device. Once prepared, translation works with no internet connection at all.

Where we are today

The translator supports 26 languagesfully on-device (with Chrome's built-in translation AI extending coverage further), streams results sentence by sentence, detects the source language locally, and keeps working offline once a language pair is cached. Everything is free and unlimited — supported by ads, never by your data. We also make LinearCN, a Chrome extension that localizes Linear.app into Chinese.

Where we're going

The same local-first architecture extends far beyond text. In rough order:

200 languages with NLLB-200

In progress

A single on-device fallback model (Meta AI’s "No Language Left Behind") covering 200 languages — from Swahili to Tamil to Icelandic.

Document translation: PDF, TXT, SRT, Markdown

Planned

Parse and translate whole files locally — subtitles keep their timestamps, documents keep their structure. Nothing is uploaded.

Image translation (on-device OCR)

Planned

Point at a screenshot, a menu photo, or a scanned page: text is recognized and translated entirely on your device.

Installable app (PWA)

Planned

The site already works offline after one visit. Next: add-to-home-screen install and an explicit "download for offline" language pack manager.

Local text-to-speech

Exploring

Hear translations spoken aloud with a neural voice that runs in your browser — no cloud TTS API.

Local speech-to-text

Exploring

Transcribe and translate audio on-device with Whisper-class models, for interviews, lectures, and voice notes.

One promise holds for every item above: it ships on-device or it doesn't ship. We will never quietly route your documents, images, or audio through a cloud API.

Questions or ideas?

Language requests move features up the queue. We read everything.

Contact us