CRED-1
Domain credibility ratings from the CRED-1 dataset: check, batch check and search news domains
At a glance
Not tested yet- GitHub stars
- 18
- Weekly downloads
- 363
- Last commit
- 2026-11-17
- Licence
- CC-BY-4.0
- Transport
- stdio
- Auth
- none
npx -y @aloth/cred1
What it does
CRED-1 is an open dataset of credibility scores for domains known to publish mis/disinformation, conspiracy theories, or other unreliable content. It ships an MCP server so AI assistants can check domain credibility directly. It is for people who want their agent to look up how a domain has been rated by public source-rating initiatives.

Tools it gives your agent
check_domain: checks a single domain and returns score, category, level, and metadata.batch_check: checks up to 100 domains at once.search_domains: searches domains by substring or regex pattern.get_stats: returns dataset statistics such as total, per-category counts, and version.get_categories: returns the category taxonomy with descriptions and score ranges.
Quick start
Install the package globally, then point your MCP host at the cred1-mcp binary.
npm install -g @aloth/cred1
Add it to Claude Code from a terminal:
claude mcp add cred-1 -- npx -y @aloth/cred1
Or add this to .mcp.json in your project root (Claude Code) or .cursor/mcp.json (Cursor):
{
"mcpServers": {
"cred-1": {
"command": "npx",
"args": [
"-y",
"@aloth/cred1"
]
}
}
}Environment variables
The registry entry declares no environment variables.
Example prompts
- Check the credibility score for infowars.com.
- Batch check these domains and tell me which ones are flagged.
- Search the dataset for domains ending in .ru.
- What categories does the dataset use, and what score ranges do they cover?
What to watch out for
- The score is a domain-level signal, not a verdict on any individual article. The sources say no single item has been fact-checked.
- A domain not found in the dataset is treated as unknown/neutral. Absence does not mean trustworthy.
- With
npx -y @aloth/cred1, npx checks npm for a newer version each time the host starts the server and installs it. A host restart can run a release you have not reviewed. Pin the version to decide when to update. - A new version is released weekly, and every release is built from the current
mainbranch, so code changes ship the same way as dataset refreshes. - Rebuilding the dataset with
enrich_dataset.pyrequires API keys.
Not tested by tools4agents yet. Read the source repository before you install it.
Similar tools
Memory
A knowledge graph that persists facts across sessions, so an agent can recall context after the conversation ends.
DBHub
Token-efficient MCP server for PostgreSQL, MySQL, MariaDB, SQL Server, Oracle, ClickHouse, SQLite
PostgreSQL
Read only SQL access with schema introspection, so an agent can answer questions about your data without guessing column names.
Supabase MCP Server
Connects AI assistants directly to Supabase projects to manage database tables, query data, run migrations, and control edge functions.