Scan, Extract & Call
Stop typing numbers manually. Point your camera at business cards, docs, or screens to extract and dial numbers instantly.
Get Scan2Call 📱Modern browser-based AI allows for multiple input methods to suit any workflow. At Staksoft, our PDFaiGen Speech-to-Text tool implements these core features using local processing:
Upload File
Support for common audio formats for batch processing.
From URL
Directly transcribe audio hosted on the web.
Live Record
Capture and transcribe voice notes instantly.
Get your audio processed locally in three easy steps without your files ever leaving your device:
1 .Upload Audio: Select a file from your device, provide a URL, or record a live voice note.
AI Processing: Our advanced Whisper-based model analyzes the audio locally in your browser using WebGPU acceleration.
Download Text: Instantly copy the transcript or export it as a TXT or JSON file with accurate timestamps.
Using @xenova/transformers, you can enable WebGPU acceleration to handle these heavy tasks with ease.
import { pipeline } from '@xenova/transformers';
/**
* Initialize Whisper Speech-to-Text with WebGPU
* Enables local processing for files, recordings, and URLs
*/
async function startTranscription(audioSource) {
const transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-tiny.en', {
device: 'webgpu',
});
const output = await transcriber(audioSource, {
chunk_length_s: 30,
stride_length_s: 5,
return_timestamps: true,
});
return output;
}
Experience the speed and privacy of our WebGPU-powered Speech to Text converter at PDFaiGen.
Staksoft.com Innovating with private-first, on-device AI solutions. Explore our full suite at staksoft.com.
LLM integration, OCR, and on-device AI engineering from Staksoft.