# RunHarvester

> RunHarvester is a paid MCP server for web search and scraping. Web data extraction and browser workflow tools for AI assistants, scoped by API key. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Paid
- **Platforms:** API, Web
- **Website:** https://docs.runharvester.com/mcp/
- **MCP Registry name:** com.runharvester/mcp
- **Categories:** web search and scraping, browser automation
- **Listing updated:** 2026-11-16

## About

RunHarvester exposes two API families through one stateless Streamable HTTP endpoint, and your API key decides which you see. A Data key shows 12 operations for creating, retrying, cancelling and batching extraction requests, plus reading results; a Browser Workflow key shows 17 more for runs, evidence, webhooks, batches and CDP sessions within its bound workspace, project and environment. To use both, configure the endpoint twice with one key of each type. Scopes are rechecked on every call, so revocation takes effect immediately.

Responses use a consistent ok/status/data/meta envelope with safe error objects. List tools default to 5 summary records, and large content is read in bounded chunks of 16 KiB by default, up to 64 KiB. Markdown is converted deterministically from stored HTML with no LLM, capped at 2 MiB, and the docs say to treat page content as untrusted. Wait helpers poll for up to 30 seconds, and request bodies are limited to 4 MiB. A connection_info tool reports the key type and scopes so you can confirm the setup.

## Features

- **Two key types:** Data keys and Browser Workflow keys expose different tools.
- **Scope-checked calls:** Each call is rechecked against the key's scopes.
- **Bounded results:** Large content arrives in 16 KiB chunks.
- **Markdown output:** HTML is converted deterministically without an LLM.
- **Wait helpers:** Poll runs and batches for up to 30 seconds.
- **Evidence access:** Browser runs return screenshot and masked HTML evidence.

## Good for

- Extract page data into structured results from an agent
- Run a browser workflow and read its evidence
- Batch many extraction requests
- Replay a failed webhook delivery
- Check the key's scopes before a long job

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http mcp https://api.runharvester.com/mcp
```

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

```sh
https://api.runharvester.com/mcp
```

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

```json
{
  "mcpServers": {
    "mcp": {
      "url": "https://api.runharvester.com/mcp"
    }
  }
}
```

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

```json
{
  "servers": {
    "mcp": {
      "type": "http",
      "url": "https://api.runharvester.com/mcp"
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "mcp": {
      "serverUrl": "https://api.runharvester.com/mcp"
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "mcp": {
      "type": "http",
      "url": "https://api.runharvester.com/mcp"
    }
  }
}
```

## Tools

- `data_create_request`: Creates a data extraction request
- `data_get_request`: Gets a data request
- `data_get_result`: Gets a request's result
- `data_create_batch`: Creates a batch of data requests
- `data_wait_request`: Waits for a request to finish
- `data_read_result_content`: Reads bounded chunks of result content
- `browser_get_context`: Gets the bound browser workflow context
- `workflow_create_run`: Creates a browser workflow run
- `workflow_get_run`: Gets a workflow run
- `workflow_read_evidence`: Reads run evidence
- `workflow_create_webhook`: Creates a webhook
- `workflow_open_cdp_session`: Opens a CDP session
- `connection_info`: Reports key type and scopes

## FAQ

### How do I authenticate?

Send an rh_live key as a Bearer token or X-API-Key; keys in URLs are rejected.

### How many tools are there?

29 API operations plus connection helpers: 12 Data and 17 Browser Workflow.

### Can I use both families?

Yes, configure the same URL twice with a key of each type.

### Is page content trustworthy?

The docs say to treat returned page content as untrusted input.

### Is RunHarvester free?

No. RunHarvester is a paid product.

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

Run this command in your terminal: claude mcp add --transport http mcp https://api.runharvester.com/mcp

### What are the best alternatives to RunHarvester?

The closest alternatives to RunHarvester are Browser Use, Scrapling MCP Server, Firecrawl MCP Server, SearXNG Search and Oya Browser.

## Alternatives

- [Browser Use](https://ai-powered.com/mcp/browser-use): Browser Use is a freemium MCP server for browser automation.
- [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.
- [Firecrawl MCP Server](https://ai-powered.com/mcp/firecrawl-mcp-server): Firecrawl MCP Server is a freemium MCP server for web search and scraping.
- [SearXNG Search](https://ai-powered.com/mcp/searxng-search): SearXNG Search is an open-source MCP server for web search and scraping.
- [Oya Browser](https://ai-powered.com/mcp/oya-browser): Oya Browser is a freemium MCP server for browser automation.
- [Mantis](https://ai-powered.com/mcp/mantis-2): Mantis is an open-source MCP server for browser automation.
- [NeuralVerge](https://ai-powered.com/mcp/neuralverge): NeuralVerge is an MCP server for web search and scraping.
- [AI-Powered directory](https://ai-powered.com/mcp/ai-powered-directory): AI-Powered directory is a free MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/runharvester
