MCP · REST API · OpenAPI · llms.txt

Agent-native AEO

The only AEO tool AI agents can actually call. No browser required. No sales call. No manual workflow.

shell — MCP install

npx @webmole/mcp --api-key=wbm_your_key

# Works with Claude Desktop, Cursor, Windsurf
# Agents call: audit_url("https://example.com")
# → structured JSON with score + fixes

Get API Key View OpenAPI Spec

Starter plan: 100 audits/day · Bearer token auth · JSON responses · No sales call

What agent-native AEO means

Every other AEO tool requires a human. You log in. You click through a dashboard. You copy the data somewhere else. That works if you have a dedicated analyst.

Agents can't log in. They can't click. They work with APIs, MCP servers, and structured files. Agent-native AEO means WebMole is built to work as a tool in that stack — not just a product for humans to review.

This matters right now because AI coding assistants (Claude, Cursor, Copilot) are increasingly writing code that touches AEO. When an agent is building a site and asks “does this page score well for AI search?”, it needs a tool it can call. WebMole is that tool.

Quick start

Three ways to integrate. Pick the one that fits your agent runtime.

shell — REST API

curl -X POST https://webmole.ai/api/v1/audit \
  -H "Authorization: Bearer wbm_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'

# → {"score": 73, "factors": [...], "fixSuggestions": [...],
#    "rateLimit": {"remaining": 99, "limit": 100}}

typescript — SDK

import { WebMole } from '@webmole/sdk'

const wm = new WebMole({ apiKey: process.env.WM_API_KEY })

const report = await wm.audit('https://example.com')
// report.score        → 73
// report.factors      → [{id, name, status, score, max}]
// report.fixSuggestions → [{instruction, codeBlock}]

// Apply fixes autonomously — no human needed
if (report.score < 70) {
  const topFix = report.fixSuggestions[0]
  await applyFix(topFix.codeBlock)
}

shell — MCP server

npx @webmole/mcp --api-key=wbm_your_key

# Registered tools:
#   audit_url(url: string) → AuditReport
#   get_factors(url: string) → Factor[]
#   get_fixes(url: string) → FixSuggestion[]

Integration points

Every surface is machine-readable. Agents don't need a human to mediate.

REST API

POST /api/v1/audit

Bearer token auth, structured JSON response with score, per-factor analysis, fix suggestions, and rate limit headers. One endpoint, one required field.

Authorization: Bearer wbm_your_key

MCP Server

npx install

Install with one command. Works out of the box with Claude Desktop, Cursor, and Windsurf. Agents call audit_url and get back the full structured report.

Compatible: Claude Desktop · Cursor · Windsurf

OpenAPI Spec

/api/v1/openapi.json

Machine-readable spec published at /api/v1/openapi.json. Auto-generate typed clients in any language. Zero competitors publish one.

OpenAPI 3.0 · JSON Schema · auto-client generation

agents.json

/.well-known/agents.json

Flow discovery at the well-known path. AI agents can self-discover capabilities, endpoints, and auth requirements without human intervention.

Well-known standard · self-describing · agent-native

Why agents choose WebMole

Every other AEO tool is built for humans in dashboards. WebMole is built for code.

Capability	WebMole	Competitors
Self-serve API key	Immediate — no approval	Contact sales
Entry price	$29/month	$499/mo or custom
Response format	Structured JSON	HTML dashboards
Rate limit headers	Every response	Not exposed
OpenAPI spec	Published at /api/v1/openapi.json	None publish one
MCP server	npx @webmole/mcp	Profound only (enterprise-gated)
agents.json	/.well-known/agents.json	None available

Machine-readable resources

Everything an agent needs is available at a stable URL. No scraping, no login wall.

llms.txtStructured plain-text summary for LLM context windows.llms-full.txtFull documentation dump — everything an agent needs to use WebMole.OpenAPI specMachine-readable API schema. Auto-generate clients in any language.agents.jsonFlow discovery endpoint. Self-describing capabilities for agent runtimes.AGENTS.mdHuman-readable guidance for coding agents integrating with WebMole.

Common questions

What is agent-native AEO?

Agent-native AEO means your AI citation monitoring tool can be used by AI agents directly — no browser, no login, no manual workflow. WebMole exposes a REST API, MCP server, and AGENTS.md so agents can audit URLs, read fix suggestions, and integrate into CI pipelines without human intervention.

How do I install the WebMole MCP server?

Run npx @webmole/mcp --api-key=wbm_your_key in any Claude Desktop, Cursor, or Cline configuration. See the quick start section for the full config block.

Does WebMole work inside Cursor?

Yes. Install the WebMole MCP server and Cursor can run AEO audits, read fix suggestions, and integrate results directly into your workflow without leaving the editor.

What is llms.txt?

llms.txt is a machine-readable file at the root of a website that helps AI systems understand what the site is and does. WebMole's llms.txt is at webmole.ai/llms.txt and includes the full API surface, pricing, and documentation links.

Start integrating in 60 seconds

API key, MCP server, OpenAPI spec — all available immediately on any paid plan. No approval. No sales call. $29/month.

Get API Key — $29/mo View OpenAPI Spec

Also available: developer docs · agency tier · agents.json