Parlai — Ai Text To Speec

Parlai — AI Text to Speech

Hear any text on the web read aloud by AI voices. Select text, right-click, and parley — everything runs locally in your browser. No servers, no accounts, no data leaves the ship.

How it works
Highlight text on any webpage, right-click, and choose "Parlai — AI Text to Speech." Your text is spoken aloud by a natural-sounding AI voice, right there in your browser. No sign-ups, no API keys, no catch.

14 AI voices
A full crew of voices at your command. Choose from 10 GPU-accelerated voices (powered by Supertonic 2) or 7 CPU voices (powered by Piper) that work on any machine. Each voice has its own character — from polished and authoritative to warm and
conversational.

4 languages
English, Spanish, Portuguese, and French. Pick a language and Parlai translates your highlighted text before speaking it — handy for navigating foreign waters.

Two engines
• GPU mode — High-quality speech synthesis using WebGPU. Fast and expressive.
• CPU mode — Lightweight fallback using WebAssembly. Runs on any vessel, no GPU required.

Playback controls
• Play, pause, skip fore and aft (10s and 30s)
• Three speeds: Slow (0.8×), Normal (1×), Fast (1.4×)
• Pitch-preserving speed changes — no chipmunk voice, even at full sail

100% local and private
All AI models are bundled aboard the extension (~1 GB). Your text never leaves your browser for speech synthesis. The only network request Parlai ever makes is for translation — when you select a non-English language and Chrome's built-in local
translator isn't available, Google Translate is used as a fallback. No data is ever sent to Parlai's servers — there are none. Your words stay on your ship.

Settings remembered per engine
Switch between CPU and GPU freely — Parlai remembers your preferred voice, speed, language, and quality settings for each engine independently. No need to reconfigure when you change course.