MCP Integration

Prepzo is the first ATS with native Model Context Protocol (MCP) support. Connect AI agents, LLMs, and automation tools directly to your hiring pipeline.

What is MCP?

The Model Context Protocol is an open standard that lets AI models interact with external tools and data sources. Instead of building custom integrations for every AI tool, MCP provides a universal interface.

With Prepzo's MCP server, any MCP-compatible AI agent can list your open jobs, search candidates, move applications through your pipeline, schedule interviews, and pull analytics - all through natural language.

Quick Start

The MCP server is available at two endpoints:

POST /api/mcp - JSON-RPC 2.0 message endpoint
GET /api/mcp/sse - SSE transport (for streaming clients)

Both require Bearer token authentication (same API key as the REST API).

Initialize

curl -X POST https://prepzo.ai/api/mcp \
  -H "Authorization: Bearer pk_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "initialize",
    "params": {
      "protocolVersion": "2024-11-05",
      "clientInfo": { "name": "my-agent", "version": "1.0" }
    }
  }'

List Available Tools

curl -X POST https://prepzo.ai/api/mcp \
  -H "Authorization: Bearer pk_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'

Available Tools

Tool	Description
`list_jobs`	List open jobs with optional status filter
`get_job`	Get detailed job info by ID
`create_job`	Create a new job posting
`list_candidates`	List candidates with optional search
`search_candidates`	Search by name, email, or tags
`create_candidate`	Add a new candidate
`list_applications`	List applications with filters
`move_application`	Move to a pipeline stage
`reject_application`	Reject with optional reason
`schedule_interview`	Schedule an interview
`get_hiring_analytics`	Get funnel metrics and time-to-hire
`search`	Unified search across all data

Call a Tool

curl -X POST https://prepzo.ai/api/mcp \
  -H "Authorization: Bearer pk_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 3,
    "method": "tools/call",
    "params": {
      "name": "search_candidates",
      "arguments": { "query": "senior engineer" }
    }
  }'

SSE Transport

For streaming MCP clients, connect to the SSE endpoint:

# Connect to SSE (returns the message endpoint URL)
curl -N -H "Authorization: Bearer pk_live_YOUR_KEY" \
  https://prepzo.ai/api/mcp/sse

# Response (SSE stream):
# event: endpoint
# data: https://prepzo.ai/api/mcp

Use with Claude Desktop

Add Prepzo to your Claude Desktop MCP config:

{
  "mcpServers": {
    "prepzo": {
      "url": "https://prepzo.ai/api/mcp/sse",
      "headers": {
        "Authorization": "Bearer pk_live_YOUR_KEY"
      }
    }
  }
}

Use with OpenAI Agents

# Python with openai SDK
from openai import OpenAI

client = OpenAI()
response = client.responses.create(
    model="gpt-4o",
    tools=[{
        "type": "mcp",
        "server_label": "prepzo",
        "server_url": "https://prepzo.ai/api/mcp/sse",
        "headers": {"Authorization": "Bearer pk_live_YOUR_KEY"}
    }],
    input="Show me all open engineering jobs and their applicant counts"
)

Why MCP?

Universal: Works with any MCP-compatible AI tool
No custom code: AI agents can discover and use tools automatically
Natural language: "Find candidates who applied for the senior role last week"
Composable: Chain multiple tools together for complex workflows
Secure: Same Bearer token auth as the REST API