Select & Explain - Explain any term in context — 100% local, no account

by
Reading dense papers in a second language means constantly tab-switching to Google. Select & Explain fixes that: highlight any term, press Alt+E, and get a context-aware explanation right next to it — it reads the surrounding paragraph, so you get what the term means in THIS paper. Unlike other AI extensions: 100% local & private (your key, your browser, zero servers), answers in any language, and a one-time $5 with your own API key — no subscription.

Add a comment

Replies

Best
Maker
📌
Hi Product Hunt! 👋 I built Select & Explain because I don't want to turn on a new page every unfamiliar term while reading papers in English (my second language). Now just select the word, hit Alt+E, and the explanation opens right next to it — context-aware, because it reads the surrounding paragraph. I can even ask follow-ups in the same card, like "give me an example." A few things I cared about: 🔒 100% local — your API key and history never leave your browser. No account, no tracking. 🌍 Answers in your language — type any language in settings. 💸 One-time $5, bring your own key (OpenAI, Claude, Gemini, and more). No subscription. Would love your feedback — especially from anyone who reads a lot in a second language. Happy to answer any questions!
How did GPT-5.6 change the ambition or scope of what you shipped?
Select & Explain was built with GPT from the ground up. As a solo/indie maker, I used GPT to design, write, and iterate on the entire extension — it let me ship a polished, production-ready tool far faster than I could alone. GPT-5.6 also shaped what the product does at runtime. It's one of the core models users can plug in with their own key, and its stronger context understanding is what let us go beyond dictionary-style definitions: the tool reads the surrounding paragraph and explains what a term means *in this specific paper*, then supports follow-up chat like "give me an example." In terms of ambition: GPT made me bolder. Without it, this would have stayed a weekend script. With it, I could aim for a private, local, context-aware reading assistant that actually holds up for researchers reading dense papers in a second language — and offer it as a one-time purchase instead of a subscription, since capable models are now affordable via API.