Raycast Relay for Cloudflare Workers

• Setup
• Usage
• Use with Cursor
• Available Models

This project provides a relay server that allows you to use Raycast AI models through an OpenAI-compatible API interface, deployed as a Cloudflare Worker.

Deployment

One-Click Deployment (Recommended)

1. Get your Raycast AI Token in the step 1 below
2. Deploy
3. Add RAYCAST_BEARER_TOKEN in Settings -> Domain & Routes -> Variables and Secrets. You can optionally add a API_KEY, ADVANCED, INCLUDE_DEPRECATED. Follow the guide below

Prerequisites

• Node.js
• Wrangler CLI
• Cloudflare account
• Raycast API credential (Bearer token)
• HTTP Debug Tool like Proxyman or Charles on macOS

Setup

1. Get your Raycast AI Token

Proxyman: Add raycast to your SSL proxying list, resend a request to AI. You will see a request like in the photo below: Copy the API key after Authorization Bearer.

2. Clone this repository

git clone https://github.com/szcharlesji/raycast-relay
cd raycast-relay

3. Install dependencies:

npm install

3. Configure your environment variables:

# Install wrangler
npm install -g wrangler

# Set your Raycast credentials as secrets
wrangler secret put RAYCAST_BEARER_TOKEN

# Optionally set an API key for authentication, if you want to use it with cursor, follow the cursor setup
wrangler secret put API_KEY

# Optionally set to filter out the advanced AI options if you don't have the subscription.
wrangler secret put ADVANCED

# Optionall set to include deprecated models
wrangler secret put INCLUDE_DEPRECATED

4. Deploy to Cloudflare Workers:

npm run deploy

Usage

Once deployed, you can use the worker as an OpenAI-compatible API endpoint:

https://your-worker-name.your-account.workers.dev/v1

API Endpoints

• GET /v1/models - List available models
• POST /v1/chat/completions - Create a chat completion
• GET /health - Health check endpoint

Authentication

If you've set an API_KEY, include it in your requests:

Authorization: Bearer your-api-key

Use with Cursor

Raycast-relay supports Cursor, but a workaround is needed since Cursor has a known issue with custom AI endpoints other than OpenAI. Thanks to Vincent's suggestions

In order to use your relayed API endpoint in cursor:

1. Generate an API key in OpenAI Platform, you just need to use it to verify it
2. Verify this key in Cursor by putting it in Cursor Settings > Models > OpenAI API Key with the default OpenAI endpoint
3. Upload your wrangler secret API_KEY by wrangler secret put API_KEY, this needs to be the same key as the OpenAI key
4. Override your OpenAI Base URL with your wrangler endpoint
5. Save it
6. Add a custom model that you can find in the /v1/models endpoint or check out Available Models
7. Done!

Available Models

Here's a list of all the model IDs:

• raycast-ray1
• raycast-ray1-mini
• openai-gpt-4.1
• openai-gpt-4.1-mini
• openai-gpt-4.1-nano
• openai-gpt-4
• openai-gpt-4-turbo
• openai-gpt-4o
• openai-gpt-4o-mini
• openai_o1-o3
• openai_o1-o4-mini
• openai_o1-o1-mini
• openai_o1-o1
• openai_o1-o3-mini
• anthropic-claude-haiku
• anthropic-claude-sonnet
• anthropic-claude-3-7-sonnet-latest
• anthropic-claude-3-7-sonnet-latest-reasoning
• anthropic-claude-opus
• anthropic-claude-sonnet-4
• anthropic-claude-opus-4
• anthropic-claude-sonnet-4-reasoning
• perplexity-sonar
• perplexity-sonar-pro
• perplexity-sonar-reasoning
• perplexity-sonar-reasoning-pro
• groq-meta-llama/llama-4-scout-17b-16e-instruct
• groq-llama-3.3-70b-versatile
• groq-llama-3.1-8b-instant
• groq-llama3-70b-8192
• together-meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo
• mistral-open-mistral-nemo
• mistral-mistral-large-latest
• mistral-mistral-small-latest
• mistral-codestral-latest
• groq-deepseek-r1-distill-llama-70b
• google-gemini-2.5-pro
• google-gemini-1.5-flash
• google-gemini-2.5-flash
• google-gemini-2.0-flash
• google-gemini-2.0-flash-thinking
• together-deepseek-ai/DeepSeek-R1
• together-deepseek-ai/DeepSeek-V3
• xai-grok-3
• xai-grok-3-mini
• xai-grok-2-latest