# Orchard

> Orchard is a freemium MCP server for files and documents. Pay-per-call web and data utilities for agents: page to Markdown, RSS, CSV and JSON. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Freemium
- **Platforms:** API
- **Website:** https://orchardshop.xyz
- **MCP Registry name:** xyz.orchardshop/orchard
- **Categories:** files and documents, web search and scraping, developer tools
- **Listing updated:** 2026-11-15

## About

Orchard describes itself as an autonomous shop that builds and runs small tools for other agents. There are no accounts or API keys: calls cost about a cent each and are paid in USDC on Base over x402, with a few free tries per day on the cheaper tools.

The catalogue splits into web tools and text or data helpers. On the web side it converts a public page or up to 20 pages to Markdown, extracts page metadata and links, crawls a site breadth-first (up to 20 pages, same host, robots.txt respected, identifying as OrchardBot/1.0), indexes a sitemap, audits up to 25 pages for broken links, missing titles and slow or heavy pages, and reads RSS or Atom feeds. Helpers validate JSON, convert between CSV and JSON, slugify titles, hash and encode text, strip Markdown and count words.

Fetchers only reach public hosts and refuse localhost and private addresses, including after redirects, and calls have time limits and size caps.

## Features

- **Page to text or Markdown:** Convert one page, or up to 20 at once, with scripts and navigation removed.
- **Site crawl and audit:** Crawl up to 20 pages as Markdown or audit 25 pages for broken links and slow pages.
- **Sitemap inventory:** Title, description and word count for up to 25 sitemap pages.
- **RSS and Atom reader:** Latest items from a public feed, 1 to 50 per call.
- **CSV and JSON conversion:** Convert in both directions, handling quoted fields.
- **Text helpers:** Slugs, hashes, encodings, word counts and Markdown stripping.
- **Public hosts only:** Private and localhost addresses are blocked.

## Good for

- Turn a documentation page into Markdown for summarising
- Audit a site's titles, descriptions and broken links
- Pull the latest items from a feed
- Convert CSV from a task into JSON rows
- Generate URL slugs for a batch of titles

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http orchard https://orchardshop.xyz/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://orchardshop.xyz/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "orchard": {
      "url": "https://orchardshop.xyz/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "orchard": {
      "type": "http",
      "url": "https://orchardshop.xyz/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "orchard": {
      "serverUrl": "https://orchardshop.xyz/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "orchard": {
      "type": "http",
      "url": "https://orchardshop.xyz/mcp"
    }
  }
}
```

## Tools

- `page-to-text`: Fetches a public web page and returns its readable text as JSON: url, status, title, char count, and clean text with scripts, styles and tags stripped. Blocks private/localhost addresses. Free: 3 calls a day per visitor; after that $0.01 per call via x402 at https://tools.orchardshop.xyz/t/page-to-text.
- `json-check`: Validates a JSON document sent as the raw request body and returns it pretty-printed. Returns valid true/false, the formatted JSON, a count of keys, and byte size. Invalid JSON returns the parser error message. Free: 3 calls a day per visitor; after that $0.01 per call via x402 at https://tools.orchardshop.xyz/t/json-check.
- `text-stats`: Send plain text as the raw POST body and get back word count, sentence count, character count, estimated reading minutes, and the top 10 keywords (stopwords removed). Free: 3 calls a day per visitor; after that $0.01 per call via x402 at https://tools.orchardshop.xyz/t/text-stats.
- `hash-encode`: POST raw text as the body and get back its byte count, sha256, sha1, md5, hex, base64 and base64url encodings. Useful for checksums, IDs and encoding a value mid-task. Free: 3 calls a day per visitor; after that $0.01 per call via x402 at https://tools.orchardshop.xyz/t/hash-encode.
- `csv-to-json`: POST raw CSV as the body (first row is headers) and get back a JSON array of row objects, plus column names and row count. Handles quoted fields, commas and newlines inside quotes, and escaped quotes. Free: 3 calls a day per visitor; after that $0.01 per call via x402 at https://tools.orchardshop.xyz/t/csv-to-json.
- `slugify`: POST titles as JSON {"titles": [...]} or as raw text with one title per line, and get back URL-safe slugs. Strips accents, lowercases, removes apostrophes and collapses other punctuation to single hyphens. Up to 500 titles per call. Free: 3 calls a day per visitor; after that $0.01 per call via x402 at https://tools.orchardshop.xyz/t/slugify.
- `json-to-csv`: POST a JSON array of flat objects (or {"rows":[...]}) and get CSV back. Columns are the union of all keys in order of first appearance. Values are quoted and escaped per RFC 4180. Nested objects are stored as JSON text. Up to 5000 rows per call. Free: 3 calls a day per visitor; after that $0.01 per call via x402 at https://tools.orchardshop.xyz/t/json-to-csv.
- `markdown-strip`: POST raw markdown as the request body and get plain text back. Removes headings, emphasis, links (keeps the label), images (keeps the alt text), list markers, blockquotes, HTML tags and code fences (keeps code contents). Useful before summarizing or indexing markdown documents. Free: 3 calls a day per visitor; after that $0.01 per call via x402 at https://tools.orchardshop.xyz/t/markdown-strip.
- `page-to-markdown`: Fetches any public http(s) page and returns clean Markdown: headings (#), paragraphs, lists, links with absolute URLs, bold/italic, and code blocks. Scripts, styles and nav/footer are removed. Output is capped at the first 50,000 characters, with a truncated flag. Blocks localhost and private addresses. Returns JSON with url, status, title, chars, truncated and markdown. Free: 3 calls a day per vi
- `page-metadata`: Fetches any public http(s) page and returns its metadata: title, meta description, canonical URL, favicon URL, all OpenGraph (og:*) properties, and every unique absolute link with its anchor text (up to 500). Blocks localhost and private addresses. Returns JSON with url, final_url, status, title, description, canonical, favicon, og, link_count and links. Free: 3 calls a day per visitor; after that
- `site-crawl-markdown`: Crawls one public website breadth-first from a start URL, staying on the same hostname, and returns each page as clean Markdown (up to 20,000 chars per page). Reads robots.txt and skips disallowed paths, fetches at most 2 pages at a time with about 300ms between requests, and identifies as OrchardBot/1.0. Blocks localhost and private addresses, including after redirects. Answers within 45 seconds;
- `sitemap-to-index`: Reads a website's sitemap.xml and returns each page's title, meta description and word count as JSON. Checks up to 25 pages per call (one level of sitemap index), with a 25-second time limit. Every page is fetched through a public-internet-only safe fetcher. Use it to audit a site's SEO basics or to get a quick content inventory before crawling. Paid only: $0.25 per call via x402 at https://tools.
- `bulk-page-to-markdown`: Converts up to 20 public web pages to clean markdown in one call. Returns each page's title and markdown body (headings, links, list items, paragraphs; scripts and styles removed), capped at 20,000 characters per page, with a 25-second time limit. Every URL is fetched through a public-internet-only safe fetcher. Use it to pull several docs or articles into one batch instead of making 20 separate c
- `site-audit`: Crawls up to 25 pages of one website (same host only) and reports the problems an SEO or QA check looks for: broken links (up to 40 outbound links spot-checked per page), missing page titles and meta descriptions, slow pages over 1.5 seconds, and pages over 500 KB. Returns a summary of issues plus per-page details as JSON. Has a 45-second time limit and polite pacing. Every URL is fetched through 
- `rss-latest`: Returns the latest items from a public RSS 2.0 or Atom feed as JSON: title, link, published date, summary (500 chars max) and guid for each item, plus feed title and link. Inputs: ?url= (public http/https feed URL, required), ?limit= (1-50, default 10). Caps: 3 MB feed read, 50 items max, 15 s per call. Private and local addresses are refused. Feeds that are HTML pages, PDFs or binary files return
- `token-check`: Returns a fact summary of a Base token: name, symbol, decimals, total supply, proxy/upgradeable status, owner and whether ownership is renounced, which risky functions exist in the contract code (mint, pause, blacklist, fee setters, ownership transfer), and the token's deepest DEX pools on Base (Uniswap v2/v3, Aerodrome) with reserves, USD price and liquidity. Fields: ok, token, abilities, market.

## FAQ

### How do I pay?

Per call in USDC on Base over x402; there are no accounts or keys.

### Can I try it free?

Yes, the cheaper tools have a few free tries per day without a wallet.

### Does the crawler respect robots.txt?

Yes, it skips disallowed paths and identifies as OrchardBot/1.0.

### Can it fetch internal addresses?

No, localhost and private addresses are blocked.

### Is Orchard free?

Orchard has a free plan, with paid plans for more usage or features.

### How do I add Orchard to Claude Code?

Run this command in your terminal: claude mcp add --transport http orchard https://orchardshop.xyz/mcp

### What are the best alternatives to Orchard?

The closest alternatives to Orchard are REA, Scrapling MCP Server, Codebase Memory, GitHub and Serena MCP: the IDE for your agent.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Scrapling MCP Server](https://ai-powered.com/mcp/scrapling-mcp-server): Scrapling MCP Server is an open-source MCP server for web search and scraping.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Firecrawl MCP Server](https://ai-powered.com/mcp/firecrawl-mcp-server): Firecrawl MCP Server is a freemium MCP server for web search and scraping.
- [Ouroboros](https://ai-powered.com/mcp/ouroboros): Ouroboros is an open-source MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/orchard-2
