☁️ Cloudflare AI Worker — Free LLM API Proxy

🚫 No need to pay for any AI API. Use this template to get 100,000 free AI requests per day powered by Cloudflare Workers AI — no credit card, no OpenAI bill, no server, no GPU needed.

---

💡 Why Use This?

Cloudflare gives you 100,000 free AI inference requests every day through their Workers AI platform. This template wraps it into a simple API so you can use it with Make.com, n8n, any app, or your own code — just like OpenAI, but completely free.

---

✨ Features

Feature	Details
🆓 100% Free	Cloudflare Workers AI free tier — no credit card needed
🔐 Auth Protected	Bearer token authentication with your own secret key
🌐 CORS Enabled	Works from any frontend, Make.com, n8n, Postman
🤖 Multiple Models	Switch between Llama, Mistral, Phi, Qwen and more
⚡ Streaming	Supports both streaming and non-streaming responses
📦 OpenAI Format	Compatible messages array format — drop-in replacement

---

🤖 Supported Models (All Free)

@cf/meta/llama-3.3-70b-instruct-fp8-fast   ← Recommended (Best Quality)
@cf/meta/llama-3-8b-instruct
@cf/meta/llama-2-7b-chat-fp16
@cf/mistral/mistral-7b-instruct-v0.2
@cf/deepseek-ai/deepseek-r1-distill-qwen-32b
@cf/qwen/qwen1.5-0.5b-chat
@cf/microsoft/phi-2

More models: Cloudflare Workers AI Models Docs

---

🚀 Step-by-Step Deployment Guide

Step 1 — Open Cloudflare Dashboard

Go to https://dash.cloudflare.com and log in (or create a free account).

In the left sidebar, find and click "Workers & Pages".

---

Step 2 — Create a New Worker

Click the "Create application" button (top right area).

You will see several options. Click "Start with Hello World!".

You will see two fields:

• Worker name → Type any name you want (e.g. my-ai-worker)
• Worker preview → Ignore this

Click "Deploy".

---

Step 3 — Add Workers AI Binding

After deploying, you'll land on your worker's overview page. You'll see a box like this:

Click on "Bindings" inside that box.

---

Step 4 — Select Workers AI

A panel will open on the left side with binding options.

Scroll down and click "Worker AI".

---

Step 5 — Set the Binding Name

After selecting Worker AI, a form appears:

In the "Variable name" field, type exactly:

AI

⚠️ Important: The name MUST be AI (capital letters). The code uses env.AI to call the models.

Click "Add binding" to save.

---

Step 6 — Add Your Secret API Key

Now go to the "Settings" tab of your Worker.

Find "Variables and Secrets" (usually the 2nd or 3rd section).

Click "Add" or "+ Add" on the right side. A form will appear:

Field	Value
Type	Secret
Variable name	API_KEY
Value	Your chosen secret password (e.g. my-super-secret-key-123)

⚠️ Important: Variable name MUST be API_KEY (exact match). The code checks env.API_KEY.

🔐 Tip: Make your API key long and random — treat it like a password. Anyone with this key can use your worker.

Click "Deploy" to save the secret.

---

Step 7 — Paste the Worker Code

Click "Edit Code" (top right button on the worker page).

In the code editor:

1. Select all existing code (Ctrl+A)
2. Delete it
3. Paste the contents of worker.js from this repository

Click "Deploy" (top right).

---

Step 8 — Get Your Worker URL 🎉

Your worker URL will be shown at the top of the page. It looks like:

https://my-ai-worker.YOUR-SUBDOMAIN.workers.dev

That's your free AI API endpoint! You're done.

---

📡 How to Use the API

Health Check (GET)

curl https://my-ai-worker.your-subdomain.workers.dev

Returns status + list of available models.

---

Chat Completion (POST)

curl -X POST https://my-ai-worker.your-subdomain.workers.dev \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "@cf/meta/llama-3.3-70b-instruct-fp8-fast",
    "messages": [
      { "role": "system", "content": "You are a helpful assistant." },
      { "role": "user", "content": "What is 2+2?" }
    ],
    "max_tokens": 512,
    "stream": false
  }'

Example Response

{
  "id": "cf-1711234567890",
  "object": "chat.completion",
  "model": "@cf/meta/llama-3.3-70b-instruct-fp8-fast",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "2+2 equals 4."
      },
      "finish_reason": "stop"
    }
  ]
}

---

🔧 Use in Make.com / n8n

Make.com (HTTP Module)

Field	Value
URL	https://my-ai-worker.your-subdomain.workers.dev
Method	POST
Headers	Authorization: Bearer YOUR_API_KEY
Body type	JSON
Body	See example above

n8n (HTTP Request Node)

Field	Value
Method	POST
URL	Your worker URL
Authentication	Header Auth → Key: Authorization, Value: Bearer YOUR_API_KEY
Body	JSON with messages array

---

📁 Project Structure

cloudflare-worker-template/
├── worker.js          ← Main worker code (paste this into Cloudflare)
├── wrangler.toml      ← Wrangler CLI config (for developers)
├── README.md          ← This file
└── screenshots/        ← Dashboard screenshots for the guide
    ├── 1.png
    ├── 2.png
    └── 3.png

---

🛠️ Switch AI Model

In worker.js, line 27:

const DEFAULT_MODEL = "@cf/meta/llama-3.3-70b-instruct-fp8-fast";

Change to any model from the supported list. Or pass model in the request body to switch per-request.

---

🔒 Security Notes

• Your API_KEY is stored as a Secret in Cloudflare — never exposed in code
• Always use Bearer token in Authorization header
• CORS is set to * — restrict it to your domain if needed

---

👨‍💻 Author

Usman Ali — @dotusmanali

---

⭐ Like this project?

Give it a star on GitHub and share it with your friends! It helps a lot 🙏