Hi HN — I built Vocab Top because I kept looking up the same words over and over and never retaining them. Dictionary definitions weren't sticking.
The idea: use generative AI to surround every word with multiple memory anchors — a custom image, a mnemonic trick, example sentences, native audio, and grammatical forms — so your brain has more to hold onto.
Tech stack: Next.js 16, React 19, Tailwind CSS. The AI generation uses Google Gemini and FriendliAI. Audio is Google Cloud TTS. Words are saved to local storage — no account required.
The newest feature is Word Forms: when you search any word, you see all its conjugations, plurals, and gender variants. Click any form to copy it or hear its pronunciation.
Design decisions I'm still wrestling with: - Local storage vs. user accounts (simplicity vs. cross-device sync) - Pattern-based prompt construction vs. fine-tuned models for mnemonic quality - How much content to generate per word (too much = overwhelming, too little = not useful)
Would appreciate feedback on the UX, the AI generation quality, or the overall approach. Especially curious if anyone has thoughts on whether visual mnemonics actually help retention or if it's just a nice-to-have