Introducing the Fabric Trust API

Trust scores are now available programmatically. Query any AI service by slug, get the composite score, all six signal breakdowns, and coverage data in a single call. Build safety checks into your agent workflows, CI/CD pipelines, or tool selection logic.

Why an API?

The Trust Index website is great for manual lookups. But if you are building an agent that selects tools at runtime, or a CI pipeline that validates dependencies on every push, you need scores delivered as structured data. That is what the API provides.

A single endpoint. One call per service. Millisecond response times. Every field you see on the Trust Index detail page is available in the JSON response.

How it works

The simplest call is a lookup by slug:

# Look up any service by slug
curl https://api.fabriclayer.ai/lookup?slug=openai \
  -H "Authorization: Bearer fl_live_..."

The response includes everything you need to make a trust decision:

{
  "slug": "openai",
  "score": 4.51,
  "status": "trusted",
  "coverage": 0.83,
  "signals": {
    "vulnerability": 4.8,
    "operational": 4.5,
    "maintenance": 5.0,
    "adoption": 3.9,
    "transparency": 4.5,
    "publisher_trust": 4.3
  }
}

The score is a weighted mean across six signals (Vulnerability at 25%, the other five at 15% each), renormalized over evaluated signals only. The coverage field tells you how many signals had data available. Coverage below 0.40 means the service is unverified.

Key features

• Single lookups by slug, name, or package identifier
• Batch queries to check multiple services in one call (Pro and Enterprise)
• Active alerts for CVEs, deprecations, ownership changes, and downtime events
• Full signal breakdown with all six dimension scores and weights
• Coverage data so you know how much evidence backs the score

Beyond REST: MCP and GitHub Action

The API is the foundation, but we have also shipped two integrations that bring trust checks closer to where decisions happen.

The Fabric MCP server lets AI assistants like Claude Desktop query trust scores during a conversation. When your agent considers calling an external tool, it can check the score first. No context switching, no manual lookup.

The Fabric Trust GitHub Action runs trust checks in CI/CD. Add it to any workflow and it scans your dependencies against the Trust Index on every push. If anything falls below your threshold, the check blocks the merge.

Pricing

The free tier includes 1,000 requests per month with no API key required. Pro starts at $49/mo with 50,000 requests, batch lookups, and alerts. Publishers can claim and verify their services for $19/mo. Enterprise at $299/mo adds SLA and dedicated support.