Clarviaclarvia

AEO Optimization Guide

AI Engine Optimization (AEO) makes your service discoverable and usable by AI agents. This guide breaks down every scoring dimension with actionable steps.

Scan your service first →
Your stack:

Agent Framework Integration

Integrate Clarvia AEO scoring directly into your multi-agent workflows. Use these examples as LangChain tools, CrewAI agents, or OpenAI function calls.

LangChain

@tool decorator

Wrap the Clarvia API as a LangChain tool. Agents can call it to evaluate any API's agent-readiness before integration.

CrewAI

@tool + Agent

Define a CrewAI tool and assign it to an API Evaluator agent that finds the most agent-friendly APIs for your crew.

OpenAI Function Calling

tools schema

Define the Clarvia API as an OpenAI-compatible function. Works with GPT-4, Claude, and any model that supports tool use.

Quick Wins (Under 2 Hours)

+3-5 ptsAdd X-RateLimit-* headers
30 min
+2-3 ptsPublish robots.txt allowing AI agents
10 min
+3-5 ptsAdd a /health endpoint
1 hour
+5-8 ptsReturn structured JSON errors
2 hours
+3-5 ptsAdd .well-known/ai-plugin.json
30 min

API Accessibility

25 pts max

How easily AI agents can discover, reach, and authenticate with your API endpoints.

Ensure public reachability

Up to 10 pts

Your API must return a 2xx response at a publicly accessible URL. Agents need to reach you without manual setup.

Optimize response speed

Up to 10 pts

Target sub-200ms p50 latency. Use CDN, connection pooling, and edge caching. Agents abandon slow APIs.

Document authentication

Up to 5 pts

Publish OpenAPI security schemes. Support Bearer tokens or API keys with clear instructions.

Data Structuring

25 pts max

How well your data is organized for machine consumption — schemas, pricing, and error handling.

Publish OpenAPI / JSON Schema

Up to 10 pts

Provide a complete OpenAPI spec with typed schemas for every endpoint. Agents parse these to understand your API.

Make pricing machine-readable

Up to 5 pts

Add a structured pricing endpoint or embed pricing metadata. Agents need to compare costs programmatically.

Return structured errors

Up to 10 pts

Use consistent JSON error format: { error: { code, message, docs_url } }. Never return HTML error pages.

Agent Compatibility

25 pts max

Direct support for AI agent protocols, discovery mechanisms, and robot policies.

Implement MCP support

Up to 10 pts

Publish MCP (Model Context Protocol) tool definitions. This is the fastest path to agent integration.

Set agent-friendly robot policies

Up to 5 pts

Allow AI agents in robots.txt. Publish .well-known/ai-plugin.json for ChatGPT/agent discovery.

Register on discovery platforms

Up to 10 pts

List your API on MCP registries, RapidAPI, or publish a .well-known/clarvia.json profile.

Trust Signals

25 pts max

Signals that help agents (and their users) trust your service — uptime, docs, and freshness.

Publish a status page

Up to 8 pts

Use Betteruptime, Instatus, or similar. Expose an API-accessible /health endpoint. Agents check before calling.

Write excellent documentation

Up to 10 pts

Include interactive examples, code snippets in 3+ languages, and a quickstart guide under 5 minutes.

Show active maintenance

Up to 7 pts

Maintain a public changelog. Keep GitHub active with recent commits. Agents prefer actively maintained APIs.

Ready to measure your progress?

Follow our Quick Wins checklist — most developers improve their score by 15+ points in one afternoon.

Scan Now — Free