PDF Reader v12 — Now Live

Read Documents With Spatial Intelligence

AI-powered read-aloud with pixel-perfect word tracking, voice cloning, and intelligent page summaries — all inside your browser.

📖 Drop PDF or Browse Instant Reader
Live Reader ↓
Scroll to explore
0
Word Lookahead
0
ms Pause Latency
0
AI Voice Presets
0
Column Detection
Core Engine

Sequential Word
Tracking Engine

Every word on the page is mapped to exact canvas coordinates. The forward-only highlight pointer follows speech synthesis in real-time — never jumping, never wrapping.

🔍
Noise & Artifact Filter
Strips headers, footers, page numbers, URLs, citation markers, and encoding artifacts before text extraction.
14 Filter Rules
📐
Two-Column Layout Detection
Automatically detects multi-column PDFs — reads left column top-to-bottom, then right. Full-width titles correctly span both.
Y-Proximity Grouping
Voice Suite

Multi-Engine
Speech Synthesis

Three synthesis engines with automatic failover. From AI-cloned voices to GPU-accelerated Bark models on free Colab runtimes.

🎤
ElevenLabs Clone
Upload a 30-second voice sample. The AI clones it and reads your documents in your own voice.
Neural Clone
🏔️
Bark Colab GPU
Run Suno Bark on free T4 GPU via Colab. 8 voice presets, emotion detection, and VRAM watchdog.
Free GPU
🌐
Web Speech API
Zero-latency browser-native synthesis with word boundary events for real-time highlight tracking.
Zero Setup
AI Intelligence

Streaming AI
Page Summaries

One click generates a concise bullet-point summary of any page. Streamed in real-time with full XSS sandboxing — all text inserted via textContent or sanitized DOM.

🛡️
Security Sandboxed
No unsanitized DOM injections. Every AI output and PDF text item passes through strict sanitization to block XSS payloads.
✨ AI Summary — Page 1
Live Reader

Start Reading
Right Now

The complete PDF Reader is embedded below. Upload a document, hit play, and watch the word-by-word highlight follow along.

narrative://pdf-reader-v12-final