OrchidNoteDon't just open PDFs.
Understand them.
A PDF reader for people who read books and papers seriously. Highlight, write down your thoughts and ask AI — annotations stay on your computer and the original file is never changed.
Deep Residual Learning for Image Recognition
Deep convolutional neural networks have led to a series of breakthroughs for image classification. When deeper networks are able to start converging, a degradation problem has been exposed: with the network depth increasing, accuracy gets saturated and then degrades rapidly.
In this paper, we address the degradation problem by introducing a deep residual learning framework. Instead of hoping each few stacked layers directly fit a desired underlying mapping, we explicitly let these layers fit a residual mapping.
⧉ Copy table| method | top-1 err. | top-5 err. |
|---|---|---|
| VGG-16 | 28.07 | 9.33 |
| ResNet-152 | 19.38 | 4.49 |
Highlight it, then keep the thought
Select text to highlight, underline or write a thought. Thoughts, AI answers and your reading log line up by page in “This book” on the right — finish the book and your notes are already in order.
- Annotations stay on your computer; the PDF isn't changed
- Export a PDF with standard annotations any reader can show
Stuck? Box it and ask
Select a sentence or draw a box around a figure and press Alt + Q; use @ to bring in a whole chapter. Tap “Background”, “Another angle” or “Ask me first” and the AI reads along with you — only up to the page you've reached.
- Bring your own API key: DeepSeek, Kimi, Qwen, OpenAI and more
- Answers cite page numbers — click to jump there
The whole book on one page
OrchidNote draws the whole book as one map from its table of contents. Chapters you haven't read stay folded; finish one and its branches grow in — the map grows as you read.
Scanned pages you can select, search and copy
Scanned pages are recognized in the background as soon as you open them. Hover over a table and copy it in one click — paste into Excel or Word with borders and merged cells intact.
- Recognition runs on your computer; nothing is uploaded
- Export the whole book to Markdown with headings, tables and images
And also
Reading
Your recent books as a wall of covers: where you left off and how long you read this week.
Translation
Translate a page or a selection — offline with a local model, or with an AI model.
Screenshot rack
Grab any part of the screen; shots hang along the top edge. Click one to ask AI, recognize text or pin it.
Pages & files
Delete, reorder, rotate, insert and extract pages; merge by dragging in File Explorer; convert Word to PDF.
Remove ads / watermarksPro
Finds promos and watermarks repeated on every page; pages that might be damaged are left as they are.
Edit & stampPro
Move, resize and delete images and shapes on a page; add headers, footers and watermarks.
FAQ
Will OrchidNote change my PDF?
No. Annotations and thoughts are stored on your computer. OrchidNote only writes a file when you export, save page changes or remove ads.
What's the difference between Free and Pro?
Free includes all reading, annotation, OCR, export, page management and AI features. Edit objects, headers / footers / watermarks and ad removal work in Free too, but saved pages carry a small line of text in the corner; Pro saves without it. See Pricing.
Does Ask AI cost extra?
It's free once you sign in, using your own API key. You pay the model provider directly; OrchidNote adds no markup and doesn't relay your requests.
Are my files uploaded?
No. Recognition and translation run on your computer. Only when you ask AI is the selected content sent — directly to the model provider you configured. See the Privacy Policy.