Skip to content
Agent Readiness Infrastructure

Can AI Agents Actually Use Your API?

Measure your API's Agent Readiness — from discovery and documentation to authentication and execution. Deterministic checks, evidence-based scores, and actionable fixes.

No signup · passive scan · results in seconds
agentbadge scan
$ agentbadge scan https://api.example.com
✓ discovery /openapi.json
✓ OpenAPI schema detected
✓ authentication documented
✗ structured error schema missing
◐ capability description inferred
AGENT READINESS
76 / 100
+8 after fix
Immediate Proof

What AgentBadge Actually Checks

142+ deterministic checks across 15 categories — evidence-based scoring, not LLM hallucination. Every result is backed by fetched proof.

142+
Automated checks
across 15 categories
4
Scoring pillars
Discovery · Understandability · Executability · Verifiability
3
Evidence levels
VERIFIED · CONFLICT · INFERRED
100%
Reproducible
deterministic, no LLM hallucination
📋
OpenAPI Analysis
Schema parsing, auth detection, error handling checks
🔍
AI-Agent Discovery
llms.txt, robots.txt, agent-guide, well-known URIs
🔌
WebMCP Support
webmcp.json validation, MCP server descriptor checks
📊
Evidence-Based Scoring
Every score backed by fetched proof, not guesswork
Conceptual Model

The Agent Interaction Pipeline

AgentBadge scores your API across four sequential stages — from discovery to execution. Each stage is independently measurable.

🔍 01

Discover

Can agents find your API?

llms.txt, robots.txt, WebMCP manifest, well-known URIs

📖 02

Understand

Can agents read your docs?

OpenAPI spec, agent-guide, structured schemas, capability descriptions

🔑 03

Access

Can agents authenticate?

API keys, OAuth flows, x402 payments, auth documentation

04

Act

Can agents execute successfully?

Error schemas, rate limits, idempotency, structured responses

The problem

Your API can be excellent and still be invisible to agents.

Humans can fill gaps in documentation with context and intuition. Agents need machine-readable signals that tell them what exists, how to authenticate, what to call and how to recover from errors.

01 / DISCOVERY

Can an agent find you?

Machine-readable discovery, well-known resources, documentation and capability signals.

02 / UNDERSTANDING

Can it understand you?

OpenAPI, endpoint descriptions, parameters, responses and machine-readable semantics.

03 / AUTHENTICATION

Can it authenticate?

Clear, structured authentication requirements instead of instructions buried in prose.

04 / ACTION

Can it act reliably?

Structured responses, errors and evidence that reduce guessing and failed calls.

SCAN Observe public API signals.
EVIDENCE Show exactly why a rule passed or failed.
SCORE One readable score with category breakdown.
FIX Automate safe fixes, assist with semantic ones.
MONITOR Catch regressions as APIs change.
Agent Passport

Verify Agent Identity

Agent Readiness tells whether an agent can use your API. Agent Passport helps establish verifiable identity — so you know who's calling.

Evidence-first

Not "AI thinks you're 76/100."

Every important finding is tied to observable evidence. The ruleset is versioned and reproducible.

$ report.json
AB-001 OpenAPI VERIFIED
GET /openapi.json → 200 confidence 1.0
Authentication VERIFIED
POST /refund CONFLICT
Guide capability INFERRED
Structured errors MISSING
ruleset: agent-readiness-v1.2
report_hash: 9a31...f02d
DETERMINISTIC BEFORE INTELLIGENT

AI is a copilot, not the judge.

Deterministic checks handle facts we can prove. AI can help interpret ambiguous documentation, propose fixes and explain findings — but inferred capabilities require human confirmation.

Don't certify. Measure.

From diagnosis to outcome

Don't just tell developers what's broken. Help them fix it.

The core product loop is Measure → Prove → Improve. Safe fixes can be automated; semantic changes stay assisted.

TYPE A / DETERMINISTIC

Safe to automate

Known, mechanically verifiable changes.

+ add /agent-guide.json
+ add machine-readable metadata
+ update badge configuration
76 → 84   +8 Guide added
TYPE B / ASSISTED

Human confirms semantics

AI proposes; the API owner decides.

INFERRED · confidence 0.71
POST /refund
"Refund a completed payment"
Developer workflow

Works where developers already work.

CLI

Run locally

Scan an API without installing an agent or changing your backend.

npx @agentbadge/cli scan URL
GITHUB ACTION

Catch regressions in CI

Use score thresholds and machine-readable reports before deployment.

uses: agentbadge/scan-action@v1
BADGE

Show measured readiness

Put the current score in your README. Measurement becomes part of your public developer presence.

Agent Ready · 91/100
Agent Knowledge Layer

The product is machine-readable too.

AI agents should not have to reverse-engineer a marketing site. AgentBadge exposes a structured knowledge layer that lets an agent discover the product, understand capabilities and choose the right action.

  • Product context and onboarding
  • Knowledge map connecting concepts and capabilities
  • Scanner and CLI instructions
  • Machine-readable concepts and reports
  • Article-specific agent knowledge endpoints
Open Agent Knowledge Layer
GET /agent-guide/ 200 OK
machine-readable index
→ /context
→ /learn
→ /knowledge-map.json
→ /concepts/scoring
→ /capabilities/scanner
→ /capabilities/cli
Designed for humans to read and agents to navigate.
Business model

Free measurement. Paid automation and continuity.

The measurement layer stays open and useful. Revenue comes from continuous monitoring, CI automation, active verification, integrations and enterprise workflows.

Free

$0 forever
  • Public passive scan
  • Score + evidence
  • CLI
  • Badge
  • Open ruleset
Scan free →
recommended

Pro

$9 / domain / mo
  • Continuous monitoring
  • Regression alerts
  • Score history + trends
  • Fresh badge / no stale state
  • CI integrations
Start monitoring →

Enterprise

Custom
  • Active verification
  • SSO / audit reports
  • Private rulesets
  • SLA / dedicated support
  • B2B API / bulk scoring
Talk to us →

Need more than a score?

Our team builds MCP servers, smart contracts, and AI agent systems.

Frequently Asked Questions

Quick answers about AgentBadge, agent readiness, and on-chain identity.

What is Agent Readiness?

Agent Readiness is the ability of your API or service to be discovered, understood, and used by an AI agent without human intervention. It extends SEO principles to machine-readable interfaces. Learn more →

Read more
What does AgentBadge measure?

AgentBadge measures 122 agent readiness rules across 18 categories — discovery, documentation, authentication, executability, and verifiability. Each check produces evidence, not opinions. Learn more →

Read more
Is OpenAPI enough?

OpenAPI is necessary but not sufficient. It covers syntax (endpoints, parameters) but not semantics, execution (auth flows), or safety (idempotency, retries). Agents need all 8 context layers. Learn more →

Read more
How does AgentBadge score?

AgentBadge scores across four pillars: Discovery (20%), Understandability (25%), Executability (30%), Verifiability (25%). Each pillar aggregates multiple categories of checks. Learn more →

Read more
What is an Agent Passport?

An Agent Passport is a non-transferable NFT on HTS that provides on-chain identity for AI agents. It includes a DID, tier (Bronze–Platinum), and self-declared capabilities.

Read more

Explore More

  • Blog

    Deep dives into agent readiness, MCP, and the agentic web

  • FAQ

    Common questions about AgentBadge and agent readiness

  • Use Cases

    Real-world scenarios for agent-ready APIs

  • What Is Agent Readiness?

    The canonical guide to agent readiness

The thesis

Measure the new interface between APIs and agents.

AgentBadge is not trying to certify the internet. It is building an open measurement standard and the installed workflow around it: scanner → evidence → fix → CI → badge → monitoring.

Learn what agent readiness means → · Read our deep dives on the blog →