Introducing the FakeVerifier Model

A cutting-edge Hugging Face-powered classification model that assesses claims, headlines, and social posts with unprecedented accuracy and speed.

Multi-Label Classification

Classifies input into Likely Real, Likely Fake, or Unverified with nuanced analysis

Lightning Fast

Provides instant verification results optimized for real-time fact-checking

Confidence Scoring

Transparent confidence metrics to help you understand the reliability of results

What It Does

Classifies input into Likely Real, Likely Fake, or Unverified

Provides a confidence score for transparency

Designed for quick, everyday fact-checking

How to Use

Get started with FakeVerifier in three simple steps

Open Verify Screen

Navigate to the verification interface and prepare your claim or text for analysis

Input Your Claim

Paste a headline, quote, social post, or any short text you want to verify

Get Results

Receive instant verdict with confidence score and detailed analysis

Example Prompts

Try these sample claims to see FakeVerifier in action

"NASA just confirmed water on Venus."

"Scientists discover a new room-temperature superconductor."

"The Moon will appear twice as big next week."

Prompt Template

Use this layout when prompting the model. Replace bracketed sections with your content.

Claim:
[Paste the exact claim, headline, or short post]

Context (optional):
[Any brief background that helps interpret the claim]

Link (optional):
[URL to the source post/article]

Output style:
- Return a verdict: Likely Real / Likely Fake / Unverified
- Include a confidence percentage (0–100%)
- Keep it concise

API Usage

Public API: coming soon

We’ll publish full integration instructions, SDK helpers, and rate-limit details here. In the meantime, the website already uses the model server-side.

Coming Soon

A multi-evidence model with richer reasoning and source suggestions for even more comprehensive verification.