Layer 01
User device / browser
The product experience begins in the user's browser.
Responsive web interfaceLocal product controls
Product case studyPublic Beta
Founder • Product Engineer
DROP2HEAR
Give your words a voice.
A privacy-first web application that transforms text, PDFs, scanned documents, images, screenshots, Word files, and other supported content into audio—helping users listen instead of read.
AngularTypeScriptOCRWeb Speech APIKokoro TTSCloudflareGitHub Actions
Product story
The problem
People often receive information in formats that require sustained reading—long PDFs, scanned documents, screenshots, Word files, and text-heavy content.
This creates friction when users
- are multitasking
- prefer listening
- are working with long documents
- need accessibility-oriented alternatives
- want to consume information while moving
Solution
Turn documents into something you can listen to.
A format-aware processing path turns different kinds of source content into a controllable listening experience.
- Step 01
Input
- Step 02
Text extraction
- Step 03
OCR when necessary
- Step 04
Content normalization
- Step 05
Speech generation
- Step 06
Interactive listening
Product architecture
A browser-first document-to-audio pipeline.
An understandable view of how document input moves through extraction, normalization, speech, playback, and optional output.
Client-side processingPrivacy-first design
Layer 02
Document input
Content enters through format-aware input paths.
TextPDFScanned PDFImagesScreenshotsDOCXTXTMarkdownLayer 03
Document processing
The appropriate extraction strategy prepares readable content.
PDF text extractionOCRDocument parsingText normalizationLayer 04
Audio engine
Normalized content is routed to an available speech option.
Browser SpeechNatural Audio / Kokoro TTSLayer 05
Playback experience
Users control how they move through and listen to the content.
PlayPauseResumeSkipDocument navigationLayer 06
Optional output
Compatible Natural Audio output can be retained for later listening.
WAV export
Privacy architecture
Privacy by design.
Drop2Hear is designed to process supported documents directly in the user's browser wherever the feature permits, reducing the need to transmit document content to a backend service.
Browser-side processing
No account required
Minimal server dependency
User-controlled documents
Processing behavior depends on the document type, selected feature, and browser capability.
Product capabilities
Features built around listening.
Format support, extraction, listening controls, and optional audio output are organized around one practical product workflow.
- Searchable PDF support
- Scanned PDF OCR
- Image OCR
- Screenshot OCR
- DOCX support
- TXT support
- Markdown support
- Browser text-to-speech
- Play / pause
- Resume
- Skip
- Long-document navigation
- Natural Audio generation where compatible
- WAV export for generated Natural Audio
- No account required
Engineering rationale
Engineering decisions behind the product.
The product architecture favors clear, practical tradeoffs that support privacy, format breadth, and incremental capability.
01
Browser-first processing
Supports privacy, responsiveness, and reduced backend dependence.
02
Multiple document pipelines
Different document formats require different extraction and parsing strategies.
03
OCR fallback
Scanned PDFs and images do not contain directly searchable text.
04
Browser speech
Provides immediate speech capability through available browser and device voices.
05
Natural Audio
Offers a higher-quality listening option where the device and feature support it.
06
Progressive product architecture
Leaves room for future language, audio, and AI capabilities without forcing them into the initial product.
Founder • Product Engineer
Built end-to-end.
Designed, engineered, and deployed end-to-end by Anwesh Kasanagottu.
- Conceived the product
- Designed the user experience
- Designed the application architecture
- Implemented the frontend
- Implemented document-processing workflows
- Integrated OCR
- Implemented speech capabilities
- Implemented Natural Audio support
- Implemented WAV export
- Created deployment automation
- Deployed the application
- Launched the public beta
DevOps
From development to production.
A compact delivery path takes product changes from local development through automated deployment.
- Development
- Git
- GitHub
- GitHub Actions
- Cloudflare
- drop2hear.com
Technology stack
A focused stack, organized by purpose.
Technology choices grouped by the part of the product they enable.
Frontend
- Angular
- TypeScript
Document processing
- PDF parsing
- DOCX parsing
- OCR
- Client-side processing
Audio
- Web Speech API
- Kokoro TTS
- WAV generation
Platform
- Cloudflare
Engineering
- Git
- GitHub
- GitHub Actions
Exploring / Planned
What’s next.
Potential additions under consideration as the product evolves. These capabilities are not presented as currently available.
Multilingual OCR
Additional speech languages
Translation + listening
AI-powered document summaries
AI-powered simple explanations of complex text
More natural voices
Podcast-style audio
AI roadmap preview
Optional intelligence, added deliberately.
- Document
- Text extraction
- Optional AI layer
- Summary / Explanation / Translation
- Speech generation
Product principles
The decisions the product should keep honoring.
- 01Privacy first
- 02Useful before complex
- 03Listen instead of read
- 04Browser where possible
- 05No account barrier
- 06Progressive enhancement