CRED-1 is an open, reproducible domain-level credibility dataset combining multiple openly-licensed source lists with computed enrichment signals. It provides credibility scores for 2,672 domains known to publish mis/disinformation, conspiracy theories, or other unreliable content.

🎓 Presented at ACM WebSci 2026 (Braunschweig). Landing page: aloth.github.io/agentic-ai-information-integrity/cred-1. First production integration: Trackless Links for iOS and macOS, with free codes for readers and attendees: gutscheinhub.de/ratgeber/trackless-links-cred-1-acm-websci-2026.

Paper: A. Loth, M. Kappes, and M.-O. Pahl, "CRED-1: An Open Multi-Signal Domain Credibility Dataset for Automated Pre-Bunking of Online Misinformation," Preprint, 2026. doi:10.2139/ssrn.6448466

Key Features

• 2,672 domains with credibility scores (0.0–1.0)
• Fully reproducible — Python pipeline rebuilds the dataset from scratch
• Multi-signal scoring combining source labels, domain age, web popularity, fact-check frequency, and threat intelligence
• Privacy-preserving — designed for on-device client-side deployment (no server calls needed)
• Two openly-licensed sources — no proprietary data dependencies

Quick Start

import json

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

domain = "infowars.com"
if domain in cred:
    score = cred[domain]["credibility_score"]  # 0.0 (least credible) to 1.0 (most credible)
    print(f"{domain}: credibility = {score}")
else:
    print(f"{domain}: not in dataset (neutral)")

Dataset Schema

JSON Format (cred1_current.json)

{
  "infowars.com": {
    "category": "fake",
    "credibility_score": 0.14,
    "domain_age_years": 26.4,
    "domain_registered": "1999-10-04T04:00:00Z",
    "iffy_factual": "VL",
    "iffy_bias": "FN",
    "iffy_score": 0.1,
    "factcheck_claims": 52,
    "safe_browsing_flagged": false,
    "score_age": 0.2,
    "score_cat": 0.05,
    "score_factcheck": 0.0,
    "score_iffy": 0.1,
    "score_safebrowsing": 0.05,
    "score_tranco": 0.1,
    "sources": 2,
    "tranco_rank": 4382
  }
}

CSV Format (cred1_current.csv)

Same fields as JSON, in tabular format with 18 columns. Sorted by credibility_score ascending (least credible first).

Compact Format (cred1_compact.json)

Minimal format for on-device embedding (e.g., browser extensions). Short keys, no whitespace, ~168KB.

Scoring Model

CRED-1 computes credibility scores as a weighted blend of five independent signals:

Remaining weight (when signals are unavailable) defaults to the source category score.

Data Sources

Reproduce the Dataset

# 1. Build base dataset (fetch + merge sources)
python3 pipeline/build_dataset.py

# 2. Enrich with signals (requires Google Cloud API key)
export GOOGLE_API_KEY="your-key-here"  # or macOS Keychain
python3 pipeline/enrich_dataset.py

# Individual enrichment steps:
python3 pipeline/enrich_dataset.py --step tranco
python3 pipeline/enrich_dataset.py --step rdap
python3 pipeline/enrich_dataset.py --step factcheck
python3 pipeline/enrich_dataset.py --step safebrowsing
python3 pipeline/enrich_dataset.py --step score

Requirements: Python 3.10+, no external dependencies (stdlib only).

Category Distribution

Applications

CRED-1 is designed for:

• Browser extensions — on-device pre-bunking at the content delivery stage
• Misinformation research — ground truth for domain-level credibility studies
• Content moderation — automated flagging of low-credibility sources
• Education — media literacy tools and curricula

Citation

If you use CRED-1 in your research, please cite the paper:

@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},
  note      = {Preprint available at SSRN}
}

To cite the dataset archive directly:

@dataset{loth2026cred1data,
  title     = {{CRED-1}: An Open Multi-Signal Domain Credibility Dataset},
  author    = {Loth, Alexander},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.18769460}
}

Contributing

Found a misclassified domain? Want to propose a new credibility signal? We welcome community input.

• 🌐 Report a Domain — flag a domain that seems misscored or missing
• 💡 Propose a Signal — suggest a new credibility signal for the pipeline
• ❓ Ask a Question — methodology, usage, or reproduction questions

License

This repository (code and data) is licensed under CC BY 4.0.

Acknowledgments

This dataset builds on the work of:

• Melissa Zimdars and the OpenSources.co project
• The Iffy.news team at the Reynolds Journalism Institute
• Google Fact Check Tools and Safe Browsing APIs

Powered by Google Fact Check Tools and Google Safe Browsing.