A pane of clarity on every frame.

Drag a rectangle on any web page. PaneTrans pins a glass popup with live translation that updates the moment the text changes — live chat, video subtitles, ticker text, anywhere.

The free, on-device tiers keep every translation inside your browser — no API key, nothing sent anywhere. Or switch on the optional Cloud subscription for top-quality LLM translation on any device.

See pricing How it works

See it actually work

Right-click → drag a rectangle → translate. The popup keeps watching as new content rolls in.

What PaneTrans actually does

Drag, then watch.

Right-click anywhere → Translate Region → drag a rectangle. A draggable popup shows the original and the translation side-by-side.

It keeps watching.

A MutationObserver tracks the rectangle. Live chat lines, subtitle changes, ticker updates — the popup retranslates automatically. Like sticking a real-time subtitle strip onto any slice of the page.

Local AI, picked per task — plus optional Cloud.

Free, on-device tier: OPUS-MT (~160 MB / pair, real-time on WebGPU) plus NLLB-200 / M2M-100. Pro adds on-device LLMs Qwen 3 / Gemma 3. New: an optional Cloud tier streams a hosted LLM for top quality on any device — no download, no GPU.

Why anyone working with foreign-language content uses this

On-device by default. The free and Pro models run from your browser cache — no third-party server, works offline. The optional Cloud tier is the only part that sends text out, and only when you choose it.
Persistent, not single-shot. Most translators give you one translation and stop. PaneTrans pins a popup that follows the region until you close it.
Bilingual YouTube subtitles as a separate, opt-in feature for video learners.
Open architecture. Built on Transformers.js + Xenova ONNX models. You can audit the model list in src/utils/languages.js.

Ready to try it?

The Free tier handles real-time translation for 10+ languages. Pro unlocks LLM-grade models for nuanced text.

Pricing