API Documentation

1Authentication

All API requests require a valid API key passed in the Authorization header as a Bearer token.

Create and manage API keys in your dashboard. Keys are hashed with SHA-256 before storage.

2Endpoint

3Query Parameters

4Response Headers

Every successful response includes these custom headers with processing metrics.

5Examples

6JSON Response Format

Send Accept: application/json to receive a structured JSON response instead of raw Markdown.

{
  "markdown": "# Article Title\n\nClean markdown content...",
  "url": "https://news.ycombinator.com/article",
  "title": "Article Title",
  "tokens": 1250,
  "originalTokens": 16180,
  "savingsPercent": 92.3,
  "cached": true,
  "fetchMs": 0,
  "format": "markdown",
  "model": "claude-sonnet-4"
}

7Batch Endpoint

Request Body

{
  "urls": [
    "https://news.ycombinator.com",
    "https://react.dev/learn",
    { "url": "https://docs.anthropic.com", "selector": "article" }
  ],
  "model": "claude-sonnet-4"
}

Each item in urls can be a plain string or an object with url and optional selector.

Example

Response

{
  "results": [
    { "url": "...", "markdown": "...", "tokens": 1250, "status": 200 },
    { "url": "...", "markdown": "...", "tokens": 890, "status": 200 }
  ],
  "total": 2,
  "success": 2,
  "failed": 0,
  "model": "claude-sonnet-4"
}

8MCP Server

Use StripFeed directly from Claude Code, Cursor, Windsurf, or any MCP-compatible client. No code needed.

{
  "mcpServers": {
    "stripfeed": {
      "command": "npx",
      "args": ["-y", "@stripfeed/mcp-server"],
      "env": {
        "STRIPFEED_API_KEY": "sf_live_your_key"
      }
    }
  }
}

The MCP server exposes two tools: fetch_url and batch_fetch. Your AI assistant can use them to read any webpage as clean Markdown. See the full setup guide on GitHub or install from npm.

9TypeScript SDK

The official TypeScript SDK provides a typed wrapper around the StripFeed API. Zero dependencies, works with Node.js 18+.

import StripFeed from "stripfeed";

const sf = new StripFeed("sf_live_your_api_key");

// Full result with metadata
const result = await sf.fetch("https://docs.anthropic.com");
console.log(result.markdown);
console.log(`Tokens: ${result.tokens} (saved ${result.savingsPercent}%)`);

// Just the Markdown string
const md = await sf.fetchMarkdown("https://docs.anthropic.com");

10Python SDK

The official Python SDK provides a simple wrapper around the StripFeed API. Zero dependencies, works with Python 3.9+.

from stripfeed import StripFeed

sf = StripFeed("sf_live_your_api_key")

# Full result with metadata
result = sf.fetch("https://docs.anthropic.com")
print(result["markdown"])
print(f"Tokens: {result['tokens']} (saved {result['savingsPercent']}%)")

# Just the Markdown string
md = sf.fetch_markdown("https://docs.anthropic.com")

11Errors

Errors return a JSON body with an error field describing the issue.

{
  "error": "Missing or invalid Authorization header"
}

12Rate Limiting

Requests are rate-limited based on your plan. There are two layers of rate limiting.

Caching

StripFeed automatically caches processed results for 1 hour (default). Cached responses are instant (0ms processing time) and still count toward your monthly quota. Check the X-StripFeed-Cache header to see if a response was served from cache.