Castor
Open Studio
⚡ Fully Offline & Local-First

Your Voice, Untethered.
Private Speech Studio.

Read articles, PDFs, and documents aloud without ever sending your data to the cloud. Fully offline synthesis powered by local ONNX transformers.

Launch Free StudioExplore Features
Active Document Reader

The Odyssey of Speech

"Human speech is a delicate mechanism of waves. In the local sandbox of the modern browser, Castor processes and synthesizes language directly, guaranteeing absolute privacy."

Line 1 of 12
Kokoro Local Engine (ONNX)
On-Device Processing
  • PDF / DOCX Ingestion (Pure JS Workers)
  • IndexedDB relational catalog metadata
  • OPFS (Origin Private File System) wav storage
Try importing a file

Engineered for Privacy & Convenience

Everything you need for an immersive, robust document listening experience, packaged in an offline-first PWA.

100% Local-First Privacy

Your private documents remain strictly private. Speech synthesis and parsing run inside secure browser sandboxes, preventing external server leaks.

Hybrid Cloud Boost

Seamlessly switch to high-performance remote models like Gemini-Lite, Voxtral, or Orpheus for premium cloud voice capabilities under user consent.

Offline-First Library

Keep your reading list stored directly on your disk via Origin Private File System (OPFS) binaries and IndexedDB relational catalogs. Listen anywhere, anytime.

Lock Screen Keep-Alive

No playback cutting off. Using custom silent loop mechanisms, Castor continues running seamlessly in the background and respects native lock screen control keys.

Private Cloud Sync

Save and restore your personal reading lists and scripts effortlessly across multiple devices with secure, encrypted cloud backup.

Premium Responsive UI

A gorgeous layout designed for mobile, tablet, and desktop viewports alike. Supports smooth animations, interactive sliders, and strict light/dark custom color themes.

How Castor Works

Castor relies on the power of modern client-side APIs to render, manage, and read text without intermediate servers.

01

Ingest & Parse

Paste raw text or upload a PDF/DOCX file. Castor's native Web Worker parsed extraction splits text into clean sentence arrays instantly.

02

Cache & Store

Scripts are saved in IndexedDB, and parsed audio bundles are saved into Origin Private File System (OPFS) as high-quality WAV buffers on your device's local drive.

03

Offline Synthesize

Castor reads sentences aloud using in-browser Kokoro pipeline transformers, or queries OpenRouter API when choosing premium remote models online.

Simple, Transparent Structure

Listen locally for free, or top up your USD wallet for premium remote model support.

Free Tier

Local Reader

$0 / forever

Synthesize speech locally on your device with no limits and absolute privacy.

  • Unlimited Kokoro 82M Local ONNX synthesis
  • Unlimited local IndexedDB/OPFS storage
  • PWA background keep-alive support
Start Listening Free
Cloud Boost
Hybrid Option

Cloud Synthesis

Pay-As-You-Go / USD wallet

Load your wallet to invoke premium remote models and voices on demand.

  • Gemini-Lite, Voxtral, and Orpheus cloud models
  • Multi-language cloud voice strategies
  • Secure cloud backup & library sync
Launch Studio & Top Up Wallet