Hindi PDF Editor

hindipdfeditor.com

Hindi PDF Editor is a privacy-first document platform for Indian languages. Pruning Labs built its in-browser WebAssembly parser with Indic glyph shaping and a specialised Devanagari OCR model, processing documents entirely client-side with no server upload and no telemetry.

The problem

Devanagari matras were mangled by cloud PDF tools, which treat Indic scripts as if glyphs map one-to-one onto code points. They do not.

Separately, enterprise and government users would not upload documents to a third-party server at all - which ruled out every cloud-based option regardless of rendering quality.

What we pruned

• Server-side uploads and document retention pruned
• Broken font rendering engines pruned
• Account walls and signup friction pruned

What we engineered

• In-browser WebAssembly parser with correct Indic glyph shaping
• Zero-telemetry client-side processing
• Specialised OCR model for Hindi administrative scripts

How it is built

• The parser and shaping engine are compiled from Rust to WebAssembly and run in the user's browser tab. The document never leaves the device.
• Glyph shaping handles conjuncts, matras, and reordering properly rather than assuming a one-to-one code-point mapping.
• OCR is trained on Hindi administrative documents specifically - the forms, notices, and records people actually need to process.
• No account, no telemetry, no retention. There is no server-side document store to breach.

Stack

• Rust
• WebAssembly
• PDF.js engine
• Devanagari OCR

Disciplines involved