Media
2 items
Description
Color-codes AI responses by evidence strength. Copy or auto-fill credibility prompts. Works on Claude, ChatGPT, and Gemini. AI says everything with the same confident tone — whether it searched for it, recalled it from training data, or made it up. This extension makes the difference visible. HOW IT WORKS 1. Add the credibility prompt to your AI's custom instructions (one-click copy from the extension popup) 2. The AI labels every claim with evidence tags like [S1], [M2+R2], [S3+F] 3. The extension reads those tags and color-codes paragraphs in real time COLOR SYSTEM 🟢 Green — Verified: strong sources, confirmed facts 🟠 Orange — Caution: weak source, possibly outdated memory, framework-dependent judgment 🔴 Red — Alert: fragile evidence, open speculation, time-sensitive but not searched TWO MODES - Simple — clean emoji badges (Verified / Caution / Alert), minimal clutter - Audit — full color backgrounds, hoverable tag explanations, fragile underlines FEATURES - Auto-highlights AI responses as they stream - One-click prompt copy from popup - Prompt version selector injected into chat toolbar - Hides raw label text for cleaner reading - Toggle Simple/Audit mode with one click WORKS ON - claude.ai - chatgpt.com - gemini.google.com NO DATA COLLECTION This extension runs entirely in your browser. No data is sent anywhere. No API calls. No tracking. Pure CSS/JS injection. Open source: https://github.com/Kunkun2116/AI-Credibility