# Anyfetch MCP server

> Enables AI agents to search the web and fetch content from sites that typically block automated scrapers using browser-grade requests.

- Repository: https://github.com/aliasmode/anyfetch
- Page: https://hostedmcp.org/servers/aliasmode/anyfetch
- Transport: stdio
- Stars: 2
- License: Apache-2.0
- Latest version: 0.8.3

## Install

### Claude Code

```sh
claude mcp add anyfetch -- npx -y @aliasmode/anyfetch
```

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

```json
{
  "mcpServers": {
    "anyfetch": {
      "command": "npx",
      "args": [
        "-y",
        "@aliasmode/anyfetch"
      ]
    }
  }
}
```

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

```json
{
  "servers": {
    "anyfetch": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@aliasmode/anyfetch"
      ]
    }
  }
}
```

### Claude Desktop (`claude_desktop_config.json`)

```json
{
  "mcpServers": {
    "anyfetch": {
      "command": "npx",
      "args": [
        "-y",
        "@aliasmode/anyfetch"
      ]
    }
  }
}
```

### Codex (`~/.codex/config.toml`)

```toml
[mcp_servers.anyfetch]
command = "npx"
args = ["-y", "@aliasmode/anyfetch"]
```

### Gemini CLI (`~/.gemini/settings.json`)

```json
{
  "mcpServers": {
    "anyfetch": {
      "command": "npx",
      "args": [
        "-y",
        "@aliasmode/anyfetch"
      ]
    }
  }
}
```

## Tools (28)

- `url`
- `query`
- `max_chars`
- `offset`
- `fresh`
- `urls`
- `provider`
- `site`
- `time`
- `page`
- `read`
- `searxng_url`
- `auto`
- `brave`
- `yahoo`
- `duckduckgo`
- `searxng`
- `github`
- `stackoverflow`
- `hackernews`
- `arxiv`
- `papers`
- `challenge_required`
- `http_error`
- `rate_limited`
- `timeout`
- `unsupported_content`
- `extraction_failed`

## FAQ

### Can Anyfetch read websites that block AI agents?

Yes, it uses browser-grade requests including Safari's TLS, HTTP/2, and header fingerprinting to bypass many bot checks and 403 errors.

### How does it handle JavaScript-heavy websites?

It uses the site's public API or feed for specific platforms like Reddit and Stack Overflow, and can optionally use Lightpanda for pages that require scripts.

### What format is the fetched content returned in?

The server returns clean Markdown, keeping links, tables, and code blocks intact.