Incognito PrivoAI

Voice-to-text, AI chat, web translation, AI reading. Gemini API key or local AI — no backend, data stays in browser.

PrivoAI — a privacy-first, all-in-one AI toolbox. ──────────────────────────────── 🆕 What's New (v28.14) ──────────────────────────────── ▶ A much better AI chat • Replies appear as they're written — no waiting for the whole answer • Clean formatting: headings, lists, tables and code are easy to read, not one wall of text • One-tap Copy or Regenerate on every reply • Prompt library: name, search and pin your go-to prompts, and import/export them • Attach up to 10 images and 5 documents at once (large PDFs included), with drag-and-drop ▶ Smarter, more faithful translation • Translations keep the original links, citations and tables — no more broken layout • Flip between the translation and the original with one tap • Better with PDFs — even two-column academic papers translate column-by-column • Works across more site layouts ▶ Voice Notes, two ways Record then transcribe in one pass (most reliable, nothing dropped), or watch text appear as you speak (real-time preview) — pick reliability or real-time. ──────────────────────────────── ✨ Key Features ──────────────────────────────── 🌐 Webpage Translation (in-place / screenshot) What Chrome's built-in translation can't do: read bilingually, preserve journal and paper layouts (including two-column PDFs), translate Kindle ebooks and comics, and translate any page with foreign-language images. One button picks the right method automatically — layout, figures, math and code stay untouched, and it translates as you scroll. Switch the target language anytime, choose which parts to translate (title / body / references…), flip between translation and original, or turn it off to restore the page. Specially tuned for complex academic sites like Nature, ScienceDirect and arXiv — titles and body text translate correctly (including CJK titles). 🎙️ Voice Notes Record or upload audio and let AI turn speech into text. Choose "Basic Recording" (most reliable, nothing dropped) or "Real-time transcription" (text appears as you speak). Auto-detects the language, handles mixed-language speech, and offers optional side-by-side translation. Handles long recordings, and you can copy the transcript or send it to the Reader. Great for meetings, interviews and multilingual notes. 💬 Incognito Chat A private chat window that leaves no record once closed. Replies are nicely formatted and stream in as they're written, with one-tap Copy and Regenerate. Includes a searchable prompt library, and you can attach images and documents. Pick from four models — from the fastest for everyday questions to the most capable for deep analysis of long documents — plus a Local model for fully offline text chat. Not sure which to pick? The default fits most cases. 🧰 Translation Workbench Paste text or upload a screenshot for text recognition (OCR) and translation, optionally with an AI summary, opened in the Reader. 📖 Reader Turns cluttered pages into a clean, readable layout with click-to-translate paragraphs. Built-in read-aloud: pick a language and press play. Adjustable font size, line spacing and background color. 🔊 AI Read-Aloud & Summary [Beta] Turn YouTube captions, audio files, webpages or pasted text into natural speech or a spoken summary — a warm, steady voice, customizable reading instructions, and WAV output. ──────────────────────────────── 🔒 Privacy by Design ──────────────────────────────── • No backend server: every AI request goes straight from your browser • No data collection: the developer never sees your input or output • Your key, your control: the Gemini API key stays on your machine — never uploaded or synced • Local AI support: text chat and translation can use Ollama / LM Studio, keeping data fully on-device ──────────────────────────────── 🌏 Multilingual Interface ──────────────────────────────── 13 interface languages: Traditional Chinese, English, Japanese, Korean, German, Spanish, Arabic, Russian, Brazilian Portuguese, Turkish, Ukrainian, Vietnamese, Czech. ──────────────────────────────── 🚀 Getting Started ──────────────────────────────── 1. Install the extension 2. Get a free Gemini API key from Google AI Studio (aistudio.google.com) 3. Click the extension icon → Settings → paste your key 4. Start using it! Tip: Want to go fully offline? Switch to "Local model" in Settings and connect Ollama or LM Studio — text chat and translation need no Gemini key.