Now accepting early access signups

Every tool has a score.
Every agent knows
who to trust.

Kerq is the trust layer for AI agent toolchains. With one API call for real-time trust scores, security, and reliability — your agents never connect to tools blindly again.

⚡ Blazing fast — 27ms avg response

Web Scraping

ScrapeMaster

✓ Certified

Code Analysis

CodeLens AI

◆ Audited

Data Pipeline

DataForge

● Scored

Search Scores See How It Works

How It Works

Trust that's earned, not claimed

Every tool listed on Kerq gets scored across 7 proprietary trust metrics based on real performance data. No pay-to-win. No self-reported metrics.

Tool Gets Listed

Any MCP server or API tool can be listed. The moment it's added, Kerq starts tracking performance automatically.

Score Updates Live

Uptime, response time, error rate, and output quality are tracked continuously. Your Kerq Score (0-100) updates in real-time.

Agents Decide

Agent developers query the Kerq API before connecting to any tool. Higher score = more agents choosing your tool = more usage.

Kerq Composite Score — 7 Contributing Metrics

▲

Uptime

Availability

% online over rolling 30d

⏱

Response Time

Speed

avg ms per request

⚠

Error Rate

Reliability

% failed calls

⟳

Consistency

Stability

latency variance & jitter

⇅

Throughput

Capacity

requests handled per second

📈

Performance Trend

Trajectory

score direction over time

↻

Recovery

Resilience

time to recover from failures

Developers who integrate our API automatically contribute to the accuracy of every score.

Trust System

Every tool gets a score. The best earn a badge.

The Kerq Score rates every AI tool from 0–100 on real performance — free and automatic. Tools that sustain excellence earn Kerq Certified™.

● Layer 1 — Free

Kerq Score (0–100)

Automatic. Every tool gets one. No payment required, no exceptions.

✓Tracks uptime, response time, error rate, and output quality continuously
✓Score updates in real-time based on actual performance data
✓Every tool gets scored—no exceptions, no payment required
✓The industry standard rating: “94 Kerq” means something
✓Color-coded tiers: green (80+), yellow (50-79), red (<50)

Kerq Score

ScrapeMaster

99.8% uptime • 142ms avg • 0.2% errors

◆ Layer 2 — Earned

Kerq Certified™

The Michelin star of AI tools. A public track record proving a tool delivers — consistently.

✦ Earned, not bought — 30 consecutive days with all Kerq metrics at 80+ unlocks the badge
✦ 12-month public performance timeline — visible to every developer and agent evaluating a tool
✦ Embeddable badge for any website, GitHub README, or documentation
✦ Automated re-verification every 30 days — no humans, no politics, no pay-to-play
✦ Score and certification are independent — a 92 Kerq tool without the badge hasn't proven consistency yet. A 92 Kerq + Certified tool has.

Kerq Score

CodeLens AI ✓ Certified

Score + Verification = complete picture

Both signals tell a story. A tool with 92 Kerq is performing great. Add Certified and agents know it's been deeply vetted too. Neither replaces the other.

Trust scores built on automated infrastructure testing and real-world developer signals — not self-reported claims.

Built For

Two sides. One trust layer.

Whether you build tools or build agents—Kerq gives you the trust infrastructure you've been missing.

⚙

For Tool Providers

List your tool and let the data speak. Higher scores mean more agents choosing you.

→List your tool and get automatically scored from day one
→See exactly why your score is what it is—uptime %, response time, error rate breakdown
→Get actionable recommendations to improve your score
→Higher score = more agents choosing your tool = more revenue
→Optional certification for additional credibility (Audited / Certified)

List Your Tool

⚛

For Agent Developers

Never blindly trust an MCP server again. Real data, real scores, real confidence.

→API-first trust checks before connecting to any tool
→Never blindly trust an MCP server again
→Search and filter by category, minimum score, verification tier
→Real-time reliability data, not just a directory listing
→One API call to know if a tool is worth connecting to

Get Your API Key

Developer Experience

One API call. Full trust picture.

Query any tool's trust score and verification status before your agent connects.

          
from langchain_kerq import KerqTrustTool

kerq = KerqTrustTool(api_key="your-api-key")

result = kerq.invoke("github-mcp-server")
# {"trust_score": 87, "tier": "certified", "score_breakdown": {...}}
from crewai_kerq import KerqTrustTool

kerq = KerqTrustTool(api_key="your-api-key")

# Add to any CrewAI agent
agent = Agent(role="Analyst", tools=[kerq])
curl https://kerq.dev/api/tools/github-mcp-server/score \
  -H "Authorization: Bearer YOUR_API_KEY"

Every tool has a score. Every agent knows who to trust.