OrchidNoteEffective: September 26, 2026 · Last updated: September 26, 2026
OrchidNote is built on excellent open-source projects. The components below are used under their respective licenses. The full license texts ship with the program in the licenses folder of the install directory, and you can also view them from the “About” window via “View third-party components and licenses”.
Libraries
| Component | Purpose | License |
|---|---|---|
| PDFium (bblanchon.PDFium packages) | PDF rendering, text extraction, table line detection | BSD-3-Clause / Apache-2.0 |
| PDFsharp | PDF writing: annotation export, merging, outlines, page editing | MIT |
| ONNX Runtime | Local model inference | MIT |
| System.Data.SQLite | Annotation and cache databases | Public Domain |
| .NET 8 runtime and libraries | Runtime environment | MIT |
| C#/WinRT（WinRT.Runtime） | Access to the OCR built into Windows | MIT |
Models (downloaded on demand, not included in the installer)
| Model | Source | License |
|---|---|---|
| PP-OCRv5 detection / recognition | PaddlePaddle (ONNX conversion by RapidAI) | Apache-2.0 |
| PP-DocLayoutV3 | PaddlePaddle (ONNX conversion by RapidAI/RapidLayout) | Apache-2.0 |
| SLANet-plus | PaddlePaddle (ONNX conversion by RapidAI/RapidTable) | Apache-2.0 |
| opus-mt-en-zh | Helsinki-NLP OPUS-MT (Jörg Tiedemann, Santhosh Thottingal, University of Helsinki; ONNX conversion by Xenova) | Apache-2.0 |
System components
Quick recognition of scanned pages uses the Windows.Media.Ocr API built into Windows. The recognition engine is provided by the operating system and is not distributed with OrchidNote; the languages it can recognize depend on the OCR language packs installed on your system.
Trademarks
Windows and Microsoft Edge are trademarks of Microsoft Corporation. DeepSeek, Kimi, Qwen (Tongyi Qianwen), Zhipu GLM, OpenAI, Gemini and other names mentioned belong to their respective owners; OrchidNote is not affiliated with or endorsed by any of these companies.