Agent optimization at the edge. Zero origin changes.
Deploy a Cloudflare Worker in front of any origin. AI agents receive optimized content. Human visitors pass through untouched.
Runs on Cloudflare's global network.
The Worker runs in 300+ data centers worldwide. Agent detection happens in under 5ms at the CDN edge, before the request ever reaches your origin server.
Optimized content is cached in Cloudflare KV for near-instant response times. Human traffic is proxied to your origin with zero overhead.
| Option | Description |
|---|---|
| Human browser | Proxied to origin unchanged |
| /llms.txt | Served from Inception Agents API (cached in KV) |
| /.well-known/agent.json | Served from Inception Agents API |
| AI agent (GPTBot, etc.) | Optimized response with JSON-LD enrichment |
Complete setup instructions.
Every code block is complete and copy-paste-ready. Total time: ~10 minutes.
Prerequisites
- Cloudflare account with Workers enabled
- Node.js 18+
- Your Inception Agents API key from the dashboard
Step 1: Install the SDK
npm install @inception-agents/cloudflare
Step 2: Create the Worker
// src/index.ts import { createInceptionHandler } from '@inception-agents/cloudflare'; export interface Env { INCEPTION_API_KEY: string; INCEPTION_CACHE: KVNamespace; } export default { async fetch( request: Request, env: Env, ctx: ExecutionContext ): Promise<Response> { const handler = createInceptionHandler({ apiKey: env.INCEPTION_API_KEY, originUrl: 'https://your-origin.com', kvCache: env.INCEPTION_CACHE, kvCacheTtl: 300, }); return handler(request, env, ctx); }, };
Step 3: Configure wrangler.toml
name = "inception-proxy" main = "src/index.ts" compatibility_date = "2026-02-01" [[kv_namespaces]] binding = "INCEPTION_CACHE" id = "your-kv-namespace-id"
Create the KV namespace:
npx wrangler kv namespace create INCEPTION_CACHE
# Copy the ID into wrangler.toml Step 4: Set Secrets
npx wrangler secret put INCEPTION_API_KEY
# Paste your API key when prompted Step 5: Deploy
npx wrangler deploy
Step 6: Verify
# Test llms.txt curl -s https://inception-proxy.your-subdomain.workers.dev/llms.txt # Test agent detection curl -s -H "User-Agent: GPTBot/1.0" https://inception-proxy.your-subdomain.workers.dev/ # Test human passthrough curl -s https://inception-proxy.your-subdomain.workers.dev/
Detection-Only Mode
If you only need agent detection metadata without content optimization:
import { createDetectionHandler } from '@inception-agents/cloudflare'; export default { async fetch( request: Request, env: { INCEPTION_API_KEY: string } ) { const detect = createDetectionHandler({ apiKey: env.INCEPTION_API_KEY }); return detect(request); // Returns: { isAgent: true, identity: "GPTBot", provider: "openai", ... } }, };
| Option | Type | Default | Description |
|---|---|---|---|
| apiKey | string | (required) | Your Inception Agents API key |
| originUrl | string | — | Origin URL to proxy non-agent traffic |
| kvCache | KVNamespace | — | KV namespace for edge caching |
| kvCacheTtl | number | 300 | Cache TTL in seconds |
| detectionThreshold | number | 0.7 | Confidence threshold for agent detection |
| enableLlmsTxt | boolean | true | Serve llms.txt files |
| enableJsonLd | boolean | true | Inject JSON-LD for agent traffic |
| enableAgentCard | boolean | false | Serve /.well-known/agent.json |
| debug | boolean | false | Log detection events |
The Worker sits in front of any HTTP origin.
Your backend stays exactly as-is.
Shopify
Set originUrl to your myshopify.com URL.
WordPress
Proxy to your WordPress hosting provider.
Salesforce
Proxy to your Salesforce Commerce instance.
Custom Backend
Any server that responds to HTTP requests.
Common issues and fixes.
Worker returns 404
Ensure originUrl is set correctly in the handler config. Check that the origin is accessible from Cloudflare's network.
KV cache not working
Verify the KV namespace ID in wrangler.toml matches your namespace. Check that the binding name matches what you pass to kvCache.
No agent traffic in dashboard
Use curl with a known agent User-Agent to test detection. Check wrangler tail for real-time logs.
Deploy in 10 minutes.
One Worker. Every AI agent optimized at the edge.