How it works
- Open the PDF in Chrome: a paper on arXiv, a course reading, a report, or a file on your computer.
- Open Onhand and ask a question about it.
- Onhand opens the PDF in its own viewer so it can mark it up, reads it, and answers. Each point is highlighted where it comes from, with a note beside it.
What you can ask
- "What's the main finding of this paper?"
- "Where do they define the loss function?"
- "Explain the equation in section 3.2." For equations, charts and diagrams, Onhand can look at the part of the page you see.
- "What does reference [14] say?" Onhand looks up the cited paper, opens it, and answers from it.
- Select a paragraph, then ask "Explain this simply."
Which PDFs it works on
- Papers and readings online: arXiv, journal sites, course pages and other PDFs you open in Chrome.
- PDFs on your computer: turn on Allow access to file URLs for Onhand in
chrome://extensions, then open the file in Chrome. - Scanned PDFs: when a page has no selectable text, Onhand marks the region of the page instead of the words.
No upload, no copy and paste
You don't upload the PDF anywhere or paste it into a chat. Onhand reads it in your browser and sends your question with the relevant text to the AI provider you chose; with Onhand Free, that goes through Onhand's server to OpenAI. Your sessions, highlights and notes stay in your browser. Privacy details
Try it
Add Onhand to Chrome, open the page you're reading, and click the
Onhand icon or press Ctrl+Shift+Space
(⌘+Shift+Space on a Mac). Onhand Free needs no account
or API key and has a daily limit; you can also sign in with your
ChatGPT plan or use your own OpenAI, Anthropic, Google Gemini or
OpenRouter key.