JevUI – Zero-shot text classifier
Classify arbitrary text without training, built for non-technical researchers
- Problem
- Non-technical researchers need to categorize open-ended participant responses, but traditional classification requires predefined labels and training data, which is slow and high-barrier. Someone used Jev to let researchers classify arbitrary content near-instantly at negligible cost, yet everyday users still lack an easy tool.
- Solution
- A lightweight web app where users paste text or import CSV, define categories in plain language (e.g., positive/negative/neutral), and get instant labels with confidence scores using zero-shot LLM classification. Supports batch processing, export, and simple stats.
- Users
- Social science researchers, market research analysts, UX researchers, non-technical teams needing quick qualitative analysis.
- MVP
- Paste text or upload CSV; define classification criteria in natural language; run zero-shot classification via LLM and show results; export labeled table; save project history.
- Revenue
- Free tier for 100 items, then subscription based on volume ($9/mo for 1000 items), or one-time purchase for self-hosted version.
- Why now
- Zero-shot classification with LLMs is reliable enough now, and API costs have dropped significantly; an indie developer can build this with minimal code.