HR System MCP Server
Enables employee information lookup, directory listing, payroll access, and time-off request management with Okta token validation.
README
HR System MCP Server
An unofficial prototype MCP server providing HR system functionality with Okta token validation. For evaluation and testing purposes only.
š Documentation
Complete documentation available in the docs/ folder:
- docs/RAILWAY_README.md - Deploy to Railway.com (3 steps) š
- docs/DOCKER_QUICK_START.md - Run locally with Docker š³
- docs/README_INTEGRATION.md - Use the deployed server š
- docs/CLAUDE.md - Developer guide & architecture š»
- docs/INDEX.md - Complete documentation index š
Overview
The HR System MCP Server provides:
- ā Employee information lookup
- ā Employee directory listing
- ā Payroll information access
- ā Time-off request management
- ā Okta OAuth 2.0 token validation for all tool calls
- ā HTTP/NDJSON streaming support (FastMCP)
- ā Ready for Railway deployment š
Authentication
This server validates Okta access tokens for all tool calls (except initialize):
- Token Source: Okta authorization server
- Validation: JWT signature, expiration, audience claims
- Authorization Header:
Authorization: Bearer <access_token>
Quick Start
# Setup
cp env.example .env
# Edit .env with your Okta credentials
# Install dependencies
pip install -r requirements.txt
# Run in HTTP mode (for Okta MCP Adapter)
python main.py --http 8001
Configuration
.env (Environment Variables)
OKTA_DOMAIN=ijtestcustom.oktapreview.com
OKTA_AUTHORIZATION_SERVER_ID=auss2fth0mcIXHzVO1d7
OKTA_AUDIENCE=
OKTA_REQUIRED_SCOPES=
# When true (default), tools/list without auth returns 401. When false, allows unauthenticated tools/list (e.g. for gateway registration).
# PROTECTED_DISCOVERY=true
Available Tools
| Tool | Description | Parameters |
|---|---|---|
get_employee |
Get employee by ID | employee_id: str |
list_employees |
List all employees | None |
get_employee_payroll |
Get payroll info | employee_id: str |
get_time_off_requests |
Get time-off requests | employee_id: str (optional) |
Usage Examples
Direct via VS Code/Copilot
# Endpoint
http://localhost:8001/mcp
# Authorization
Authorization: Bearer <okta_access_token>
Via Okta MCP Adapter Gateway
# Gateway will:
# 1. Receive request from client
# 2. Validate Okta token
# 3. Forward to HR System MCP
# 4. Attach authorization header
Implementation Details
- Framework: FastMCP 3.0.0b1
- Server: Uvicorn (async HTTP)
- Protocol: MCP (Model Context Protocol) with NDJSON streaming
- Token Validation: JWKS-based JWT validation with signature verification
- Caching: JWKS keys cached with TTL
Request Flow
Client Request
ā
Authorization Header (Okta token)
ā
Initialize (no token needed)
ā
tools/list (validate token)
ā
tools/call (validate token)
ā
Response
š Deployment Options
Vercel (Serverless) ā”
Deploy as serverless function - automatic scaling, pay-per-use
- ā Best for: Sporadic usage, automatic scale-to-zero
- ā Free tier: 100GB bandwidth/month
- ā ļø Constraint: 10-second timeout (free), 5-min (Pro)
- š Guide: docs/VERCEL_README.md
Railway.com (Traditional Server) š
Deploy as long-running server - always-on, unlimited timeout
- ā Best for: Constant traffic, persistent connections
- ā Free tier: 500 hours/month ($5/month after)
- ā No timeout: Unlimited request duration
- š Guide: docs/RAILWAY_README.md
Docker (Local Development) š³
Run locally with Docker - full control, testing
- š Guide: docs/DOCKER_QUICK_START.md
docker-compose up -d
Recommendation:
- Use Vercel for sporadic/unpredictable usage (cheaper, auto-scales)
- Use Railway for constant traffic or if you need long timeouts
Troubleshooting
See docs/RAILWAY_DEPLOYMENT.md for complete troubleshooting guide.
Quick fixes:
- Token validation fails: Verify
OKTA_DOMAINandOKTA_AUTHORIZATION_SERVER_IDin.env - Port already in use: Change port in startup command:
python main.py --http 8002 - Missing environment variables: Copy
.envexample and fill in values - JWKS fetch error: Verify Okta domain and authorization server ID are correct
Project Structure
hr-mcp-server/
āāā main.py # FastMCP server with HTTP handler
āāā requirements.txt # Python dependencies
āāā Dockerfile # Docker container definition
āāā docker-compose.yml # Docker Compose configuration
āāā railway.json # Railway deployment config
āāā deploy-railway.sh # Deployment helper script
āāā test_server.sh # Server test script
āāā auth/ # Authentication module
ā āāā __init__.py
ā āāā okta_validator.py # Okta token validation
āāā docs/ # Documentation
āāā INDEX.md # Documentation index
āāā RAILWAY_README.md # Railway quick start
āāā RAILWAY_DEPLOYMENT.md # Complete deployment guide
āāā DOCKER_QUICK_START.md # Docker reference
āāā README_INTEGRATION.md # Usage guide
āāā CLAUDE_CODE_SETUP.md # Claude Code setup
āāā CLAUDE.md # Developer documentation
āāā ...more docs
See docs/INDEX.md for complete documentation guide.
Testing
# Using curl with Okta token
curl -X POST http://localhost:8001/mcp \
-H "Authorization: Bearer <your_okta_token>" \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/list",
"params": {}
}'
š Documentation
For complete documentation, see the docs/ folder:
- Getting Started - Documentation index
- Deploy to Railway - Cloud deployment guide
- Run with Docker - Local development
- Integration Guide - How to use the server
- Developer Guide - Architecture & development
References
Status
ā ļø Unofficial Prototype - For evaluation and testing only. Not for production use.
License: Apache 2.0
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.