v1.19.0
atlasia-ma/MoulSotv1.19.0Sep 9, 2026by abimaelmartell
Upgrading?
AI Summary
This release focuses on refining text extraction accuracy by preserving visual styling like stroke-simulated bold and horizontal scaling, while improving how symbol metadata is handled and exposed across the Rust and Node.js bindings.
Key Highlights
- Preservation of stroke-simulated bold text
- Support for horizontal text scaling
- Recovery of stale single-byte font mappings
- Exposing legacy symbol-rewrite metadata in Rust and Node
Breaking Changes
- Rust `TextItem` literals now require `legacy_symbol_rewrite: false` for items without symbol cleanup
- Node.js `legacySymbolRewrite` field is now optional and only emitted when a character was actually changed
New Features
- Exposure of legacy symbol-rewrite metadata
- Improved separation of separately clipped text runs
Full Release Notes
Changes since **1.18.0**: - Preserve stroke-simulated bold. [#511](https://github.com/firecrawl/pdf-inspector/pull/511) - Honor horizontal text scaling. [#510](https://github.com/firecrawl/pdf-inspector/pull/510) - Recover stale single-byte font mappings. [#516](https://github.com/firecrawl/pdf-inspector/pull/516) - Expose legacy symbol-rewrite metadata in Rust and Node. [#519](https://github.com/firecrawl/pdf-inspector/pull/519) - Keep separately clipped text runs apart. [#518](https://github.com/firecrawl/pdf-inspector/pull/518) **Rust migration:** `TextItem` literals need `legacy_symbol_rewrite: false` for items without symbol cleanup. The Node `legacySymbolRewrite` field is optional and emitted only when cleanup changed a character. [Full changelog](https://github.com/firecrawl/pdf-inspector/blob/v1.19.0/CHANGELOG.md) · [Compare with 1.18.0](https://github.com/firecrawl/pdf-inspector/compare/v1.18.0...v1.19.0)
Related Repositories
rasbt/LLMs-from-scratch★ 107,095
Implement a ChatGPT-like LLM in PyTorch from scratch, step by step
lyogavin/airllm★ 40,274
AirLLM 2.8T LLM inference with single 4GB GPU
NirDiamant/RAG_Techniques★ 29,967
This repository showcases various advanced techniques for Retrieval-Augmented Generation (RAG) systems. Each technique has a detailed notebook tutorial.
microsoft/OmniParser★ 25,540
A simple screen parsing tool towards pure vision based GUI agent
karpathy/nn-zero-to-hero★ 24,911
Neural Networks: Zero to Hero