# Verol

> Verol is a Chrome browser extension that connects to AI chat interfaces (ChatGPT, Claude, Gemini) and performs real-time hallucination detection and fact-checking on AI-generated responses.

Verol intercepts AI chat streams as they are being generated, extracts verifiable factual claims using a lightweight LLM, and runs them through a dedicated multi-stage verification pipeline (web search, GitHub, official documentation, live data) before injecting green/red inline annotations directly into the chat UI — without any manual prompting from the user.

## Key Facts

- **Product type:** Chrome Extension (Manifest V3)
- **Supported AI chats:** ChatGPT, Claude, Gemini (more in progress)
- **Core capability:** Real-time, passive, automatic fact-checking of AI hallucinations
- **How it works:** A 4-stage autonomous pipeline — Semantic Extraction → Targeted Grounding → Cross-Reference → UI Injection
- **Pricing:** Credit-based. Free tier available. Pro plans for power users.
- **Website:** https://verol.app
- **Install:** Chrome Web Store (search "Verol")
- **Company:** Verol Research

## Pages

- [Homepage](https://verol.app): Product overview, use cases, methodology, comparison vs native AI search, pricing
- [Blog](https://verol.app/blog): Technical articles about AI hallucinations and verification
- [Why LLMs Hallucinate](https://verol.app/blog/why-llms-hallucinate): Deep-dive article on the root causes of LLM hallucinations
- [Privacy Policy](https://verol.app/privacy): Data handling and privacy commitments
- [Terms of Service](https://verol.app/terms): Usage terms

## Use Cases

- Developers verifying AI-generated code claims (library versions, API signatures, GitHub stats)
- Researchers fact-checking AI summaries of papers and news
- Writers using AI for content who need accurate sourcing
- Anyone who relies on AI chat for decisions and wants a second opinion automatically