> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tryhuntr.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Huntr MCP

> Connect Huntr to Cursor, Claude, Codex, and other MCP clients.

Huntr exposes an **HTTP MCP server** at `https://api.tryhuntr.com/mcp`. Once connected, your AI assistant gets the **full Huntr API** as tools — research, list building, enrichment, and page operations — not just `/research`.

## All MCP tools

Every paid REST endpoint (except signup webhooks) is available as an MCP tool:

| MCP tool                  | REST equivalent                 | What it does                                                            |
| ------------------------- | ------------------------------- | ----------------------------------------------------------------------- |
| `research`                | `POST /research`                | Agent-led research — account briefs, buyer research, structured answers |
| `company_search`          | `POST /company-search`          | Build company lists with structured filters                             |
| `company_search_count`    | `POST /company-search-count`    | Preview list size (free)                                                |
| `person_search`           | `POST /person-search`           | Find people by title, employer, skills, location                        |
| `person_search_count`     | `POST /person-search-count`     | Preview people match count (free)                                       |
| `person_email`            | `POST /person-email`            | Find a work email for one person                                        |
| `person_linkedin`         | `POST /person-linkedin`         | Full LinkedIn profile from URL                                          |
| `person_linkedin_url`     | `POST /person-linkedin-url`     | Resolve LinkedIn profile URL from name + company                        |
| `linkedin_company`        | `POST /company-linkedin`        | Company profile from LinkedIn URL                                       |
| `company_linkedin_url`    | `POST /company-linkedin-url`    | Resolve LinkedIn company URL from domain                                |
| `company_linkedin_posts`  | `POST /company-linkedin-posts`  | Token-paginated posts from a LinkedIn company page                      |
| `linkedin_post`           | `POST /linkedin-post`           | Detailed data for one LinkedIn post                                     |
| `linkedin_post_reactions` | `POST /linkedin-post-reactions` | People who reacted to a LinkedIn post                                   |
| `linkedin_post_comments`  | `POST /linkedin-post-comments`  | Token-paginated comments from a LinkedIn post                           |
| `tech_stack`              | `POST /company-tech-stack`      | Detect technologies on a domain                                         |
| `company_contact`         | `POST /company-contact`         | Emails, phones, socials from company website                            |
| `job_postings`            | `POST /company-jobs`            | Active job listings                                                     |
| `find_pages`              | `POST /page-find`               | Discover relevant URLs on a domain                                      |
| `page_scrape`             | `POST /page-scrape`             | Scrape structured content from one URL                                  |
| `page_analyze`            | `POST /page-analyze`            | LLM analysis of one page                                                |
| `get_pricing`             | `GET /pricing`                  | Full pricing matrix (no charge)                                         |
| `get_balance`             | `GET /balance`                  | Remaining credits (no charge)                                           |
| `get_usage`               | `GET /usage`                    | Recent request history (no charge)                                      |

<Tip>
  Agents should call **direct tools** (`person_email`, `company_search`, `tech_stack`, …) for known lookups. Use `research` only when the task needs cross-source synthesis on **one** target.
</Tip>

## Cursor

1. Open **Cursor → Settings → MCP → Add new global MCP server** (edits `~/.cursor/mcp.json`)
2. Add:

```json theme={null}
{
  "mcpServers": {
    "huntr": {
      "type": "http",
      "url": "https://api.tryhuntr.com/mcp",
      "headers": { "x-api-key": "hntr_YOUR_KEY" }
    }
  }
}
```

3. Restart Cursor. Use **Agent mode**.

## Claude Desktop

To connect Claude Desktop using its local MCP configuration, use the third-party [`mcp-remote`](https://github.com/geelen/mcp-remote) bridge to send Huntr's API-key header. The bridge runs locally through `npx`.

<Warning>
  `mcp-remote` describes itself as an experimental proof of concept. It is not maintained by Huntr or Anthropic. Review the project before installing it.
</Warning>

<Note>
  Install **Node.js 18 or later** before continuing. This requirement applies to the Claude Desktop bridge, not to Huntr's remote MCP server. Cursor, Claude Code, and Codex can connect to Huntr directly without this bridge.
</Note>

Edit `~/Library/Application Support/Claude/claude_desktop_config.json` (macOS) or `%APPDATA%\Claude\claude_desktop_config.json` (Windows):

```json theme={null}
{
  "mcpServers": {
    "huntr": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote@latest",
        "https://api.tryhuntr.com/mcp",
        "--header",
        "x-api-key:${HUNTR_API_KEY}",
        "--transport",
        "http-only"
      ],
      "env": {
        "HUNTR_API_KEY": "hntr_YOUR_KEY"
      }
    }
  }
}
```

Fully quit and restart Claude Desktop, then confirm the Huntr tools appear. Keep your API key private and do not commit this configuration file to source control.

## Claude Code (CLI)

```bash theme={null}
claude mcp add huntr --transport http https://api.tryhuntr.com/mcp \
  -H "x-api-key: hntr_YOUR_KEY"
```

Run `/mcp` in a session to verify connection.

## Codex

Set your Huntr API key in the environment:

```bash theme={null}
export HUNTR_API_KEY="hntr_YOUR_KEY"
```

Then add to `~/.codex/config.toml`:

```toml theme={null}
[mcp_servers.huntr]
url = "https://api.tryhuntr.com/mcp"
env_http_headers = { "x-api-key" = "HUNTR_API_KEY" }
```

Restart Codex, then use `/mcp` to verify the connection.

## Example prompts

**Research (one target)**

* "Use Huntr to research Notion's enterprise motion and who owns security buying decisions."

**Search + enrich (multi-step)**

* "Search for VP Engineering contacts at linear.app, then find a work email for the best match."
* "Preview how many US SaaS companies have 50–200 employees, then fetch the first page."

**Direct enrichment**

* "Get stripe.com's tech stack and open job postings using Huntr."
* "Scrape [https://linear.app/pricing](https://linear.app/pricing) and summarize the pricing tiers."

**Meta**

* "What's my Huntr balance and the price of a standard research request?"

## Tips

* Call `get_pricing` before paid tools so the agent picks the right tier and knows costs.
* For bulk lists, chain `company_search_count` → `company_search` → `person_search` → `person_email` — do **not** loop `research`.
* Get your API key from the [dashboard](https://tryhuntr.com/dashboard).

## Next step

* [Huntr Skills](/build-with-ai/huntr-skills) — ready-made GTM workflows powered by these MCP tools
* [Research Agent guide](/guides/research-agent)
* [Choose your workflow](/concepts/choose-your-workflow)
* [API Reference](/api-reference)