PeopleSoft MCP Server
A Model Context Protocol (MCP) server that enables AI assistants to query and understand PeopleSoft HCM databases. It provides semantic tools for HR, Payroll, Benefits, Performance, and PeopleTools metadata, allowing natural language questions to be answered with accurate SQL queries.
README
PeopleSoft MCP Server
A Model Context Protocol (MCP) server that enables AI assistants to query and understand PeopleSoft HCM databases. This server provides semantic tools for HR, Payroll, Benefits, Performance, and PeopleTools metadata - allowing natural language questions to be answered with accurate SQL queries.
Features
- 41 semantic tools covering all major PeopleSoft HCM modules
- 4 documentation resources for PeopleSoft concepts and query patterns
- Direct database access via Oracle thin client (no JDBC required)
- PeopleTools introspection for understanding system architecture
- Effective dating support built into all queries
Quick Start
Prerequisites
- Python 3.11+
- Oracle Database connectivity to a PeopleSoft HCM 9.2 instance
- uv package manager (recommended)
Installation
# Clone the repository
git clone <repo-url>
cd peoplesoft-mcp
# Install dependencies
uv sync
Configuration
- Copy the example environment file and add your credentials:
cp .env.example .env
- Edit
.envwith your database credentials:
ORACLE_DSN=hostname:port/service_name
ORACLE_USER=your_username
ORACLE_PASSWORD=your_password
- Edit
.cursor/mcp.jsonand update the path to your installation.
Running the Server
uv run peoplesoft_server.py
Cursor IDE Integration
The MCP config (.cursor/mcp.json) should look like:
{
"mcpServers": {
"peoplesoft": {
"command": "uv",
"args": [
"--directory",
"/path/to/mcp_ps/",
"run",
"peoplesoft_server.py"
]
}
}
}
Credentials are loaded from .env automatically - no need to include them in the MCP config.
Available Tools
Schema Introspection (5 tools)
| Tool | Description |
|---|---|
describe_table |
Get table structure, fields, and indexes |
list_tables |
Search for tables by name pattern |
get_translate_values |
Decode field codes (XLAT values) |
get_table_indexes |
View index definitions for performance |
get_table_relationships |
Find foreign key relationships |
HR Module (5 tools)
| Tool | Description |
|---|---|
get_employee |
Get employee details by EMPLID |
search_employees |
Search employees by name, department, etc. |
get_job_history |
View job history for an employee |
get_org_chart |
Get organizational hierarchy |
get_department_info |
Get department details and headcount |
Payroll Module (5 tools)
| Tool | Description |
|---|---|
get_payroll_results |
View payroll calculation results |
get_payroll_status |
Check payroll processing status |
get_accumulator_balances |
Get YTD/MTD balances |
get_payment_info |
Get payment details |
list_calendar_runs |
List payroll calendar runs |
Benefits Module (4 tools)
| Tool | Description |
|---|---|
get_benefit_elections |
View benefit plan elections |
get_dependents |
Get dependent information |
get_beneficiaries |
Get beneficiary designations |
get_benefit_costs |
Calculate benefit costs |
Performance Module (3 tools)
| Tool | Description |
|---|---|
get_performance_reviews |
List performance documents |
get_review_details |
Get detailed review information |
search_reviews |
Search reviews by criteria |
PeopleTools Module (18 tools)
| Tool | Description |
|---|---|
get_record_definition |
Full record structure with fields and keys |
search_records |
Find records by name or description |
get_component_structure |
Component pages and navigation |
get_page_fields |
Fields on a page with record bindings |
get_peoplecode |
Find PeopleCode on records/fields |
get_permission_list_details |
Security access for permission lists |
get_roles_for_permission_list |
Roles containing a permission list |
get_process_definition |
Process Scheduler job definitions |
get_application_engine_steps |
AE program structure |
get_integration_broker_services |
IB service operations |
get_message_definition |
IB message structure |
get_query_definition |
PS Query records and fields |
get_sql_definition |
Get SQL text by SQLID (views, App Engine, PeopleCode) |
search_sql_definitions |
Search SQL objects by text |
search_peoplecode |
Search text within PeopleCode |
get_field_usage |
Impact analysis - where a field is used |
get_translate_field_values |
All XLAT values for a field |
explain_peoplesoft_concept |
Explains effective dating, SetID, etc. |
Direct Query (1 tool)
| Tool | Description |
|---|---|
query_peoplesoft_db |
Execute custom SQL queries |
Available Resources
| Resource URI | Description |
|---|---|
peoplesoft://schema-guide |
Major tables by module |
peoplesoft://concepts |
Effective dating, EMPLID, SetID, XLAT |
peoplesoft://query-examples |
SQL query patterns |
peoplesoft://peopletools-guide |
PeopleTools architecture guide |
Project Structure
peoplesoft-mcp/
├── peoplesoft_server.py # Main MCP server entry point
├── db.py # Database connection management
├── tools/ # Semantic tool modules
│ ├── introspection.py # Schema discovery tools
│ ├── hr.py # HR module tools
│ ├── payroll.py # Payroll module tools
│ ├── benefits.py # Benefits module tools
│ ├── performance.py # ePerformance tools
│ └── peopletools.py # PeopleTools metadata tools
├── tests/ # Test suites
│ ├── test_business_questions.py # HR business scenarios
│ └── test_peopletools_questions.py # Technical consultant scenarios
├── docs/ # Documentation resources
│ ├── peoplesoft_concepts.md
│ ├── peoplesoft_schema_guide.md
│ ├── peopletools_guide.md
│ ├── sql_query_examples.md
│ ├── peopletools-tables-by-tool.md # Tables required per tool
│ └── migration-analysis.md # Migration planning notes
└── pyproject.toml # Project configuration
Running Tests
# Set environment variables
export ORACLE_DSN="hostname:port/service_name"
export ORACLE_USER="username"
export ORACLE_PASSWORD="password"
# Run all tests
uv run pytest tests/ -v -s
# Run specific test suite
uv run pytest tests/test_business_questions.py -v -s
uv run pytest tests/test_peopletools_questions.py -v -s
Example Queries
The MCP enables natural language questions like:
Business Questions:
- "How many active employees are in each company?"
- "What is the average salary by department?"
- "Who reports to manager X?"
- "Show me employees without a supervisor assigned"
Technical Questions:
- "What fields does PS_JOB have?"
- "Where is the DEPTID field used?"
- "What PeopleCode runs on the JOB record?"
- "What SQL does HR_ABSV_JOB_EFFDT use?"
- "Search for SQL objects referencing PS_ABSV_REQUEST"
- "Explain effective dating in PeopleSoft"
Development
Adding New Tools
- Create a new module in
tools/or add to existing module - Define async functions that use
db.execute_query() - Add a
register_tools(mcp)function - Import and register in
peoplesoft_server.py
Key Concepts
- Effective Dating: Most PeopleSoft tables use EFFDT/EFFSEQ for history
- SetID: Controls data sharing across business units
- Translate Values: Short codes decoded via PSXLATITEM
- EMPLID/EMPL_RCD: Employee ID + employment record number
License
MIT
Changelog
v0.2.1 (2026-03-02)
- Added
get_sql_definition- fetch SQL text by SQLID from PSSQLTEXTDEFN - Added
search_sql_definitions- search SQL objects by text - PeopleTools compatibility: adjusted PSPNLGROUP, PSPNLFIELD, PSAEAPPLDEFN, PSAESTEPDEFN queries for varying column names across PeopleTools versions
- Added
docs/peopletools-tables-by-tool.md- tables required per tool
v0.2.0 (2026-02-28)
- Added modular tool architecture with 39 semantic tools
- Added PeopleTools introspection module (16 tools)
- Added comprehensive test suites (42 tests)
- Replaced connection pooling with direct connections for reliability
- Added 4 documentation resources
- Improved effective dating patterns in all queries
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。