The Trust Layer for the Internet
Evida analyses any image, link or article through multiple verification methods and gives you an explainable trust score — with the evidence to back it up.
1 free check — no signup required.
Built on the same standards trusted by newsrooms and platforms
How Evida works
Three steps. Five verification methods. One transparent score.
Submit
Upload an image, paste a link, or drop in text you want to verify.
Analyze
Five independent verification methods run in parallel — from AI detection to source reputation.
Trust Score
Get an Evida Score from 0–100 with the individual signals and human-readable evidence.
Our 5 Verification Methods
Each score is the combination of five independent checks. Here is what runs under the hood.
Metadata & Provenance
We read C2PA Content Credentials and EXIF metadata to trace where a file came from, which device or model produced it, and how it was edited. When provenance is signed, we surface the full chain of custody.
AI-Generation Detection
We check for SynthID and other invisible watermarks embedded by generative models, then run pixel-level analysis for the artifacts diffusion and GAN pipelines tend to leave behind. Multiple detectors vote together to reduce false positives.
Source & Domain Reputation
For links, we score the publisher against domain age, hosting history, and known misinformation and fact-checking databases. A brand-new domain impersonating a newsroom is treated very differently from an established outlet.
Reverse Search & Cross-Referencing
We look for the same image or article elsewhere on the web to see if it has been reused in a different context, mislabelled, or debunked before. Older appearances of the same asset are strong evidence against a fresh claim.
Consistency & Manipulation Analysis
We inspect lighting, shadows, perspective, compression artifacts, and noise patterns to spot regions that have likely been edited or spliced. Inconsistencies at the pixel level often reveal manipulation the eye can miss.
Our Commitment to Transparency
No detection tool is 100% accurate — including ours. Every Evida Score is probabilistic evidence, not an absolute verdict. Detectors can be fooled, watermarks can be stripped, and adversaries evolve. That is why we publish our methodology openly, show you every signal that fed the score, and encourage you to combine our findings with your own judgment before you decide to trust or share a piece of content.
Now in public beta — free to try, no signup required
Simple, honest pricing
Start free. Upgrade when you need more verification volume.
5 checks per month
Unlimited checks & PDF reports
Teams & API access
Frequently asked questions
Honest answers about what Evida can and cannot do.