harfbuzz/harfbuzz
Harfbuzz
HarfBuzz is a text shaping engine that converts Unicode text into correctly positioned glyphs for any writing system. It processes OpenType layout features, substitution rules, and positioning tables to produce shaped glyph sequences from font data, and also handles Apple Advanced Typography font tables. The library provides a stable C API that abstracts over platform-specific font access and rendering backends, and includes a font table caching layer to avoid redundant parsing during repeated shaping operations.
The project implements script-specific shaping logic for complex writing systems including Indic, Arabic, Hangul, and Thai, and applies the Unicode Bidirectional Algorithm to reorder glyphs for mixed left-to-right and right-to-left scripts. It manages glyph input and output as a mutable buffer, allowing incremental shaping passes and reordering without intermediate allocations. HarfBuzz also includes a font subsetting capability that creates a new font by copying only the required glyphs and tables from the source, pruning unused data to minimize file size.
The library connects with FreeType, Cairo, CoreText, and DirectWrite to handle text layout and drawing across operating systems. It produces bitmap, vector, or GPU-ready representations of glyph shapes from font outlines, and can encode glyph outlines into a compact format for direct rendering on graphics hardware. Additional capabilities include generating scalable vector glyph outlines in formats such as SVG while preserving color font data, and outputting shaped text as a bitmap or vector image for preview or integration.
Features
- Text Shaping Engines - Converts Unicode text into correctly positioned glyphs for any writing system using OpenType and Apple Advanced Typography font tables.
- Glyph Buffer Managers - Provides a mutable glyph buffer for incremental shaping passes and reordering without intermediate allocations.
- Text Shaping Integration APIs - Ships a stable C API abstracting over FreeType, CoreText, and DirectWrite for cross-platform font access.
- Platform Renderer Integrations - Connects with FreeType, Cairo, CoreText, and DirectWrite for native text rendering across platforms.
- Rendering Library Integrations - Connects with FreeType, Cairo, CoreText, and DirectWrite for text layout and drawing across operating systems.
- OpenType Table Encoding - Processes Unicode text through OpenType GSUB and GPOS tables for glyph substitution and positioning.
- Layout Rule Processors - Processes OpenType layout features, substitution rules, and positioning tables to produce shaped glyph sequences from font data.
- Rendering Backend Adapters - Integrates with FreeType, Cairo, CoreText, and DirectWrite to handle text shaping and rendering across operating systems.
- Rendering Backend Integrations - Integrates with FreeType, Cairo, CoreText, and DirectWrite to handle text shaping and drawing across operating systems.
- Font Table Caches - Caches parsed OpenType and Apple Advanced Typography font tables to avoid redundant parsing during repeated shaping.
- Bidirectional Text Layouts - Applies the Unicode Bidirectional Algorithm to reorder glyphs for mixed left-to-right and right-to-left scripts.
- Glyph Rasterization - Produces bitmap, vector, or GPU-ready representations of glyph shapes from font outlines.
- Font Subsetting Engines - Creates a new font by copying only required glyphs and tables, pruning unused data to minimize file size.
- Font Manipulation - Engine for complex text shaping.
- Networking and Utilities - Text shaping and glyph positioning library.
- Media & Communication - Text shaping engine
- Open-source projects made by Iranian developers - Listed in the “Open-source projects made by Iranian developers” section of the Github Do Not Ban Us awesome list.
Star history
How this analysis was created: This summary and feature list were written by an AI model that read the project's README and public documentation pages. Each feature links to the documentation it came from; stars, license and language come straight from the GitHub API. The model does not read the source code, and the analysis is refreshed when the project is re-analysed. Learn more on our About page.
