Preview access and installation
Email support@pagenarrator.com about preview access or installation help. Testers with repository access can follow the build and installation steps. PageNarrator is free and needs no account or activation.
What your computer needs
Use a desktop Chromium 120+ browser. Kokoro supports WebGPU and CPU-only WebAssembly. A compatible GPU generates speech faster on our tested devices; CPU setup and synthesis take longer.
Experimental OmniVoice requires WebGPU with fp16 support. Its separate first-use model download is about 1.35 GB. Missing Kokoro voice and processing files also need an initial download.
Memory use and speed depend on the voice, processing mode and document. Minimum and recommended hardware tiers are still being validated; development-PC timings do not guarantee performance on your device.
Preparing a voice
Use an internet connection when you first choose Prepare voice or Read page. Missing voice files download from Hugging Face before speech is generated on your device. Wait for preparation to finish. The same voice and processing mode can reuse cached files offline afterward. Changing voice or mode, or clearing the browser cache, may need another download. If downloading fails, check your connection and retry.
What to include when asking for help
Describe what you tried, what happened, your Chrome version, and whether PageNarrator reports GPU or CPU-only mode. Do not email private text or document URLs.
Email support@pagenarrator.com.