CRED-1: 2,672 domains scored for credibility

An open, reproducible domain credibility dataset for automated pre-bunking of online misinformation. Multi-signal scoring, CC BY 4.0 licensed.

📦 Download Dataset 📄 Read the Paper GitHub

Get started in 3 lines

Load the dataset and query any domain's credibility score.

 Python

import json

with open("data/cred1_current.json") as f:

    cred = json.load(f)

domain = "infowars.com"

score = cred[domain]["credibility_score"]  # 0.14

print(f"Credibility: {score:.2f}")  # 0.0 = least, 1.0 = most

 R

library(jsonlite)

cred <- fromJSON("data/cred1_current.json")

cred[["infowars.com"]]$credibility_score  # 0.14

Six credibility signals

Each domain is scored across multiple independent dimensions, then combined into a weighted composite.

🏷️

Source Category

Classification from curated lists (fake, conspiracy, satire, bias, etc.)

📊

Iffy News Index

Factual reporting level and political bias ratings from the Iffy Index

🔍

Fact-Check Claims

Number of fact-checked claims associated with the domain via ClaimReview

📅

Domain Age

WHOIS registration age as a trust signal (newer domains score lower)

🌐

Web Popularity

Tranco ranking as a proxy for reach and potential impact

🛡️

Safe Browsing

Google Safe Browsing threat assessment for malware and social engineering

Dataset schema

credibility_score

Composite score 0.0 - 1.0

🛡️ Used in production

CRED-1 powers source credibility warnings in Trackless Links, a Safari extension for iPhone, iPad & Mac.

🔗 Trackless Links — Safari Extension

📝 Cite this dataset

@article{loth2026cred1, title = {CRED-1: An Open Multi-Signal Domain Credibility Dataset for Automated Pre-Bunking of Online Misinformation}, author = {Loth, Alexander and Kappes, Martin and Pahl, Marc-Oliver}, year = {2026}, doi = {10.2139/ssrn.6448466}, url = {https://ssrn.com/abstract=6448466} }