Clarvia

The AEO (Agent Engine Optimization) Standard for AI Agent Tools

Score, discover, and validate 15,400+ AI agent tools for quality, accessibility, and agent-readiness.

Website · npm · API Docs · Leaderboard

Install (one command)

npx -y clarvia-mcp-server

Or add to any MCP client config:

{
  "mcpServers": {
    "clarvia": {
      "command": "npx",
      "args": ["-y", "clarvia-mcp-server"]
    }
  }
}

No install needed? Use the remote endpoint directly:

https://clarvia-api.onrender.com/mcp/

What is Clarvia?

Clarvia assigns an AEO Score (0-100) to any AI tool, API, or MCP server — measuring how easily AI agents can discover, integrate, and trust it. Think of it as "Lighthouse for AI agent tools."

4 scoring dimensions (25 points each):

Plus an optional Onchain Bonus (+25) for web3 services.

For Agents

Before calling any external tool, check if it's agent-ready.

Clarvia gives your agent a quality gate for tool selection:

Agent workflow:
  Need a payment API?
  -> search_services("payment API")
  -> clarvia_gate_check(top_result)
  -> PASS? Use it. FAIL? clarvia_find_alternatives("payment")
  -> After use: clarvia_submit_feedback(result)

Agent Grades:

• AGENT_NATIVE (80+) — Built for agents, full MCP/OpenAPI support
• AGENT_FRIENDLY (60-79) — Good API, minor gaps
• AGENT_POSSIBLE (40-59) — Usable with effort
• AGENT_HOSTILE (0-39) — Avoid for autonomous use

For Tool Makers

See your tool the way AI agents see it.

• Your AEO score — how agent-ready your tool is today
• Specific gaps — missing OpenAPI spec? No MCP server? Poor error messages?
• Category ranking — how you compare to alternatives
• Actionable fixes — each dimension tells you exactly what to improve

Quick Start by Framework

{
  "mcpServers": {
    "clarvia": {
      "command": "npx",
      "args": ["-y", "clarvia-mcp-server"]
    }
  }
}

claude mcp add clarvia -- npx -y clarvia-mcp-server

{
  "mcpServers": {
    "clarvia": {
      "command": "npx",
      "args": ["-y", "clarvia-mcp-server"]
    }
  }
}

{
  "mcpServers": {
    "clarvia": {
      "command": "npx",
      "args": ["-y", "clarvia-mcp-server"]
    }
  }
}

{
  "mcpServers": {
    "clarvia": {
      "command": "npx",
      "args": ["-y", "clarvia-mcp-server"]
    }
  }
}

{
  "experimental": {
    "modelContextProtocolServers": [
      {
        "transport": {
          "type": "stdio",
          "command": "npx",
          "args": ["-y", "clarvia-mcp-server"]
        }
      }
    ]
  }
}

https://clarvia-api.onrender.com/mcp/

See examples/ for ready-to-copy config files.

MCP Tools — Node.js Server (16 tools)

Discovery & Search

Scanning & Evaluation

Agent Safety & Gating

Setup Management

Feedback & Support

Python MCP Server (24 tools)

The backend also runs a Python-based MCP server with extended capabilities:

Access via remote endpoint: https://clarvia-api.onrender.com/mcp/

Claude Code Skills

Pre-built skills for Claude Code that use Clarvia MCP tools:

Quick setup:

claude mcp add clarvia -- npx -y clarvia-mcp-server
cp .claude/skills/clarvia-*.md /your/project/.claude/skills/

See SKILLS.md for full installation and usage details.

REST API (110+ endpoints)

Full OpenAPI spec: /openapi.json

Key endpoints:

Rate limits: 10 scans/hour (free) · 100/hour (with X-API-Key)

Agent Discovery Endpoints

Clarvia is designed to be discovered by AI agents:

Architecture

scanner/
  backend/           # FastAPI (Python 3.12+) — 110+ API endpoints + MCP server
    app/
      checks/        # 13 scoring sub-factors
      routes/        # REST API endpoints
      services/      # Supabase, PDF generation, enrichment
      mcp_server.py  # Python MCP server (24 tools)
  frontend/          # Next.js + Tailwind — clarvia.art
  mcp-server/        # Node.js MCP server (TypeScript, 16 tools)
  cli/               # CLI scanner tool
  examples/          # Framework integration configs
  github-action/     # GitHub Actions for CI/CD AEO checks

Development

# Backend
cd backend && pip install -r requirements.txt
uvicorn app.main:app --reload --port 8000

# Frontend
cd frontend && npm install && npm run dev

# MCP Server
cd mcp-server && npm install && npm run dev

# Docker (full stack)
docker compose up --build

Integrations

Links

• Website: clarvia.art
• npm: clarvia-mcp-server
• PyPI: clarvia-langchain
• API: clarvia-api.onrender.com
• MCP Registry: io.github.digitamaz/clarvia
• Remote MCP: https://clarvia-api.onrender.com/mcp/

License

MIT