dnacore-check-mcp
v0.2.6
Published
MCP server for DNACORE: check AI-readiness (llms.txt, agents.json, robots.txt, JSON-LD) and submit to the AI-native directory.
Maintainers
Readme
dnacore-check-mcp
When someone asks ChatGPT, Claude, or Perplexity to recommend a tool in your field — will your site come up?
If your site isn't AI-readable, AI assistants can't find it, cite it, or recommend it — no matter how good it is. DNACORE is the directory that fixes that. This MCP server lets your AI assistant handle the entire process for you.
Just say: "Submit my site to the AI directory."
Your AI checks your site's AI-readiness, submits it to DNACORE, and gets you listed. You never open a terminal.
Setup
Add to your MCP config:
{
"mcpServers": {
"dnacore": {
"command": "npx",
"args": ["-y", "dnacore-check-mcp"]
}
}
}Works with Claude Code (~/.claude.json), Cursor (.cursor/mcp.json), Windsurf, and any MCP-compatible AI client.
What your AI does for you
One sentence → fully listed
"Submit my site https://example.com to the AI directory"
- AI calls
submit_to_dnacore - Checks your site's AI-readiness score automatically
- Submits to DNACORE
- Returns your listing ID and edit token
You do nothing except ask.
Check without submitting
"Is my site AI-readable? What's missing?"
AI calls check_ai_readiness and tells you exactly what to fix.
Tools
submit_to_dnacore
Check AI-readiness and submit to DNACORE in one step.
Input: { "url": "https://example.com" }
Returns: submission_id, edit_token (shown only once — save it), reciprocal badge HTML, and next steps.
check_ai_readiness
Check a website's AI-readiness score (0–100) without submitting.
Input: { "url": "https://example.com" }
Returns:
{
"score": 85,
"grade": "A",
"breakdown": {
"llms_txt": { "found": true, "score": 25, "detail": "Found (4 sections)" },
"agents_json": { "found": true, "score": 25, "detail": "Valid (endpoints present)" },
"robots_txt": { "found": true, "score": 35, "detail": "All major AI crawlers allowed" },
"json_ld": { "found": false, "score": 0, "detail": "Not found" }
},
"recommendations": ["Add JSON-LD structured data — see https://schema.org"]
}Scoring
| Check | Max | What it tests |
|---|---|---|
| llms.txt | 25 pts | AI-readable site index |
| agents.json | 25 pts | Agent capability declaration |
| robots.txt | 35 pts | AI crawler permissions |
| JSON-LD | 15 pts | Schema.org structured data |
Grades: A+ (≥90) / A (≥80) / B+ (≥70) / B (≥60) / C (≥50) / D (<50)
Score ≥ 60 qualifies for DNACORE listing and trust badge.
Why this matters
AI search is replacing traditional search. ChatGPT, Perplexity, and AI coding assistants like Cursor and Claude Code browse the web on behalf of their users. They need structured signals to find and trust your site — llms.txt, agents.json, and open AI crawler permissions. Without these, your site is invisible to the AI layer.
DNACORE is the directory built for this layer. Once you're listed, AI agents worldwide can discover, cite, and recommend your site via API.
Links
- Directory: dnacore.ai
- CLI tool: dnacore-check
- Protocol: github.com/dnacore-ai/dnacore-protocol
License
MIT
