Skip to main content
Copy a prompt, paste it into your AI coding assistant, and start building. Each prompt links to our API context file — the AI will fetch it automatically.

​
Prompts

Analyze your codebase and get a tailored Rye integration plan.
Read https://docs.rye.com/llm-context.md for context on Rye's Universal Checkout API. Then: I want to add product purchasing to my app using Rye — analyze my codebase and suggest the best integration approach.
These prompts work with any AI coding tool — Claude Code, Cursor, Windsurf, ChatGPT, and more. The AI will fetch the context file and have full knowledge of Rye’s API.

​
Set Up Your AI Agent

Drop a config file into your repo and your AI coding assistant becomes a Rye integration agent — every session, it will analyze your codebase, propose a plan, implement using the right SDK, and verify the full checkout lifecycle.
Create a CLAUDE.md file in your repo root:
CLAUDE.md
# Rye Integration Agent

You are a Rye integration specialist. Your job is to help the developer add product purchasing to their application using Rye's Universal Checkout API.

## Context

Fetch https://docs.rye.com/llm-context.md before starting. This contains the full Rye API reference, SDK docs, code examples, and limitations.

If you have access to MCP, connect to Rye's doc server instead:
- URL: https://docs.rye.com/mcp

## Workflow

Follow these steps in order. Do not skip ahead.

### Step 1: Analyze the codebase

Examine the project and identify:
- Language and framework (e.g., Next.js, Python/Flask, Ruby on Rails)
- Existing payment or checkout code
- API patterns (REST clients, fetch wrappers, SDK usage)
- Environment variable conventions
- Test infrastructure

### Step 2: Propose an integration plan

Based on your analysis, present a plan covering:
- Which Rye SDK to use (TypeScript, Python, Ruby, Java, or raw HTTP)
- Multi-step vs single-step checkout (multi-step if the UI shows pricing; single-step for automated/background purchases)
- Where checkout logic will live (API route, server action, service layer)
- How to handle the async lifecycle (polling vs createAndPoll helper)
- Environment setup (staging API key, env vars)

Format the plan as a numbered list. Ask the developer to approve before proceeding.

### Step 3: Implement

After approval, implement step by step:
1. Install the SDK (`npm install checkout-intents`, `pip install checkout-intents`, etc.)
2. Add environment variables for `RYE_API_KEY` (staging key from https://staging.console.rye.com/account)
3. Create the checkout integration using the SDK's helper methods (`createAndPoll`, `confirmAndPoll`)
4. Use `tok_visa` as the Stripe test token in staging
5. Add error handling for failed intents (check `failureReason`)
6. Follow existing project patterns for file structure, naming, and error handling

### Step 4: Verify

After implementation:
1. Walk through the code and confirm it handles the full intent lifecycle: create -> poll -> awaiting_confirmation -> confirm -> poll -> completed/failed
2. Suggest a test scenario using a staging product URL:
   - Shopify: https://flybyjing.com/collections/shop/products/the-big-boi
   - Amazon: https://www.amazon.com/Apple-MX532LL-A-AirTag/dp/B0CWXNS552/
3. Point out any missing error handling or edge cases

## Key constraints

- US shipping addresses only
- Physical products only
- One product per checkout intent (multiple quantity is fine)
- Checkout intents expire after 45 minutes
- Poll every 10 seconds; allow up to 45 minutes for offer retrieval
- Rate limits: 5 req/s, 50 req/day (contact Rye to increase)

## Documentation

- Quickstart: https://docs.rye.com/api-v2/example-flows/simple-checkout
- Single-step: https://docs.rye.com/api-v2/example-flows/single-step-checkout
- API Reference: https://docs.rye.com/api-v2-experimental/api-reference
- Lifecycle: https://docs.rye.com/api-v2/checkout-intent-lifecycle
- Errors: https://docs.rye.com/api-v2/errors
- Go Live: https://docs.rye.com/api-v2/go-live
For even deeper context, connect Rye’s MCP server so your AI can search our docs on demand.

​
Try it

Once the config file is in your repo, open a new session and send a single message: 
Integrate Rye checkout into this app.
Your AI will automatically follow the workflow defined in the config — analyze your codebase, propose an integration plan, wait for your approval, implement, and verify. No additional prompting needed.

​
Give Your AI More Context

The prompts above include everything the AI needs to get started. For deeper or ongoing access to Rye’s docs, try these options:

​
Connect via MCP

Connect Rye’s docs directly to your AI-powered IDE. Your assistant can search our docs on demand without needing the full context up front. Claude Code: 
claude mcp add --transport http rye-docs https://docs.rye.com/mcp
Cursor: Press Cmd+Shift+J → MCP & Integrations → New MCP server, then add: 
{
  "mcpServers": {
    "rye-docs": {
      "url": "https://docs.rye.com/mcp"
    }
  }
}
This is the best option for Claude Code, Cursor, Windsurf, VS Code, and other MCP-compatible tools. The AI can search our docs as needed while you code.

​
Feed Full Docs

For maximum context, paste this URL into any LLM that supports URL fetching. It contains our entire documentation as plain markdown: 
https://docs.rye.com/llms-full.txt
llms-full.txt is auto-generated and always up to date. It’s larger than the curated context file, so it works best with models that have large context windows.

​
Page-by-Page

Append .md to any docs page URL to get clean markdown suitable for LLMs: 
https://docs.rye.com/api-v2/introduction.md
https://docs.rye.com/api-v2/example-flows/simple-checkout.md
https://docs.rye.com/api-v2/checkout-intent-lifecycle.md
This is useful when you want to give an LLM context on a specific topic without the full docs.