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

# Cutshort MCP Server: AI-Powered Hiring Tools for Agents

> The Cutshort MCP server exposes candidate search, job management, and sourcing as tools that AI agents like Claude and GPT-4 can call directly.

The Cutshort MCP (Model Context Protocol) server lets you integrate Cutshort's hiring capabilities into AI agents, LLM-powered assistants, and tools like Claude Desktop and Cursor. Instead of building custom API glue, you configure the MCP server once and your AI agent can search candidates, post jobs, and generate shortlists as natural tool calls.

## What MCP enables

<CardGroup cols={2}>
  <Card title="AI Hiring Assistant" icon="robot">
    Let Claude or GPT-4 search candidates and answer hiring queries
  </Card>

  <Card title="Automated Sourcing" icon="gears">
    Build agents that run end-to-end sourcing without code
  </Card>

  <Card title="Slack/Teams Bots" icon="message">
    Create HR bots that answer "who should we interview?" in real time
  </Card>

  <Card title="Custom Agents" icon="code">
    Integrate Cutshort tools into any MCP-compatible AI framework
  </Card>
</CardGroup>

## How it works

The Cutshort MCP server sits between your AI agent and the Cutshort platform, translating natural-language tool calls into structured API requests:

1. Your AI agent (Claude, GPT-4, etc.) is connected to the Cutshort MCP server.
2. The agent calls Cutshort tools (`search_candidates`, `create_job`, etc.) using natural language.
3. The MCP server translates the call to a Cutshort API request and returns results.
4. The agent uses the results to answer questions or take further actions.

## Available tools

| Tool                    | What it does                                                             |
| ----------------------- | ------------------------------------------------------------------------ |
| `search_candidates`     | Search for candidates by skills, experience, location, and quality score |
| `get_candidate_profile` | Retrieve full details for a specific candidate by ID                     |
| `create_job`            | Create a new job posting with requirements and preferences               |
| `get_shortlist`         | Fetch the AI-generated candidate shortlist for a job                     |
| `send_invite`           | Send outreach invites to one or more candidates for a job                |

## REST API vs MCP

<Tabs>
  <Tab title="REST API">
    The REST API is best for **custom integrations, ATS sync, job board tools, and scheduled data pulls**. You write the integration code — defining endpoints, handling authentication, parsing responses, and managing retries. Choose the REST API when you need precise control over request structure or want to embed Cutshort data in your own backend pipelines.
  </Tab>

  <Tab title="MCP Server">
    The MCP server is best for **AI agents, LLM assistants, and conversational HR bots**. The AI writes the calls based on natural language prompts — you don't need to specify endpoints or parse JSON by hand. Choose MCP when you want to give an agent like Claude or GPT-4 direct access to Cutshort's hiring workflows.
  </Tab>
</Tabs>

<Note>
  The MCP server uses your Cutshort API key for authentication — the same key you use for REST API calls.
</Note>