Cursor + Wayfern

The Agent reads the live web next to your code.

~/.cursor/mcp.json
OAuth with PKCE
The Cursor CLI reads the same file

Cursor's Agent knows your code. It does not know what a library's docs say this week, or how your production site looks to a visitor.

With Wayfern in mcp.json, the Agent calls scrape_page, search_web and capture_screenshot. Each call runs in an anti-detect browser in Wayfern's cloud and returns markdown, JSON or an image URL.

acme-shop

TScheckout.ts cart.ts

5 export function checkout(cart: Cart) {

6 return pay.createCheckout({

7 amount: cart.total, items: cart.lines,

8 currency: 'usd',

9 returnUrl: SITE + '/done', successUrl: SITE + '/done',

10 });

11 }

Agent Fix the checkout wrapper

Fix lib/checkout.ts to match the createCheckout docs.

Thinking…

scrape_page wayfern

url: docs.northwind.dev/sdk/v4/checkout

## createCheckout(options)

- items (was amount)

- successUrl (was returnUrl)

markdown · credits_charged: 1

checkout.ts +2 -2

The v4 docs renamed two options. Both lines now match.

Setup

One entry in mcp.json.

Put it in ~/.cursor/mcp.json for all projects, or in .cursor/mcp.json to share it with the team.

Add the server

~/.cursor/mcp.json

{
  "mcpServers": {
    "wayfern": { "url": "https://api.wayfern.com/mcp" }
  }
}

2
Sign in

Cursor starts the OAuth sign-in. Sign in on wayfern.com and click Allow.
3
Check the tools

Open Customize, MCPs. Wayfern shows its tools.

Ask the Agent

prompt

Read the current docs for the payments SDK's createCheckout call and fix our wrapper in lib/checkout.ts to match.

Other ways to connect

With a token

Set WAYFERN_API_TOKEN to a token with the mcp scope in your shell startup file, for example ~/.zshrc.

~/.cursor/mcp.json

{
  "mcpServers": {
    "wayfern": {
      "url": "https://api.wayfern.com/mcp",
      "headers": { "Authorization": "Bearer ${env:WAYFERN_API_TOKEN}" }
    }
  }
}

Cursor CLI

The agent command reads the same mcp.json.

bash

agent mcp login wayfern
agent -p --trust --approve-mcps "With Wayfern, read the changelog of our UI library and list the breaking changes"

As an agent skill

Cursor loads skills from ~/.cursor/skills and ~/.agents/skills, and also from ~/.claude/skills. The skill needs WAYFERN_API_TOKEN in the environment.
bash

mkdir -p ~/.cursor/skills/wayfern curl -fsSL https://wayfern.com/skills/wayfern/SKILL.md -o ~/.cursor/skills/wayfern/SKILL.md

Route

MCP keeps the terminal free.

With MCP, the Agent calls Wayfern as a tool and needs no terminal approval or network access in the sandbox. One sign-in covers every tool.

Remote MCP OAuth sign-in: Best fit; OAuth with PKCE, from an entry in mcp.json.
MCP + token Authorization header: Works; A wf_ token in headers, read from an environment variable.
Agent skill SKILL.md: Works; Put SKILL.md in ~/.cursor/skills/wayfern. The Agent runs curl in the terminal.
REST API api.wayfern.com/api/v1: Works; The Agent can write code that calls api.wayfern.com/api/v1 directly.
Browser API Playwright over CDP: Works; The Agent can write Playwright code that connects to a Wayfern browser over CDP.

Example tasks

Web jobs for the Agent

01

Match the docs

“Read the current docs for the payments SDK's createCheckout call and fix our wrapper in lib/checkout.ts to match.”

scrape_page at 1 credit a page.
02

Check production on a phone width

“Capture https://example.com/pricing in dark mode at 390 pixels wide and tell me if the plan cards overflow.”

capture_screenshot costs 5 credits. width goes down to 320.
03

Copy design tokens

“Get the styleguide of stripe.com and write its colours and radii as Tailwind theme tokens.”

site_styleguide costs 10 credits for a site.
04

Debug from search

“Search Google and DuckDuckGo for this hydration error and list the causes that people confirm.”

search_web bills 2 credits for each provider page.

Tools

What Cursor can call

The same tools on every route, at the REST API rates. Every result reports credits_charged.

All tool inputs

scrape_page Fetch one page through an anti-detect browser and return clean markdown, HTML, links and metadata. The screenshot, pdf and json formats add +4 credits (once); summary adds +2. 1 credit / page
extract_data Extract structured data from one or more pages with a prompt and an optional JSON Schema. Re-running over pages already fetched is free. 1 credit / uncached source page
map_site List a site's URLs from its sitemaps and page links. 2 credits / 10 URLs
capture_screenshot Capture a clean image of a page: cookie banners and ads removed, lazy images loaded, tall pages captured end to end. Returns a short-lived download URL and an artifact key, not image bytes. Slices of a full page cost nothing extra; a fail_if_contains match is not charged. 5 credits / capture
brand_profile Identity, colours, logos, social links and company metadata for a domain. 10 credits / profile
site_styleguide Colours, typography, spacing, radii, shadows and the design tokens a site declares. 10 credits / site
site_fonts Font families, where each is served from, available weights, file formats and where they are used. 5 credits / site
crawl_site Start a bounded multi-page crawl. Returns a job id immediately and charges nothing up front; pages are billed as the crawl runs. 1 credit / successful page
crawl_status Poll a crawl started with crawl_site and read a page of its results. The pages were billed as they were crawled. Free
search_web Search Google, Bing, and DuckDuckGo. Returns ranked results, ads, provider failures, pagination, and related queries. 2 credits / provider page
list_search_engines List live provider ids and labels. Free

Good to know

Approvals: The Agent asks before each MCP tool call by default. In Cursor 3.6 and later, Auto-review runs allowlisted tools at once.
localhost: Wayfern's browsers run in Wayfern's cloud, so they cannot reach localhost on your computer. Use a public staging URL.