Unofficial HPO MCP Server
Provides access to the Human Phenotype Ontology (HPO) API with tools for searching, browsing hierarchy, and analyzing phenotypic terms used in genetic research.
README

Unofficial HPO MCP Server
An MCP (Model Context Protocol) server that provides access to the Human Phenotype Ontology (HPO) API. The HPO contains over 18,000 terms describing human phenotypic abnormalities and is widely used in genetic research and clinical diagnostics.
Developed by Augmented Nature
Features
This MCP server provides 12 comprehensive tools to interact with the HPO API:
Core Search & Information Tools
- search_hpo_terms - Search for HPO terms by keyword, ID, or synonym with pagination support
- get_hpo_term - Get detailed information about a specific HPO term by its ID
- get_all_hpo_terms - List all HPO terms with pagination
- batch_get_hpo_terms - Retrieve multiple HPO terms in a single request (up to 20 terms)
Hierarchical Navigation Tools
- get_hpo_ancestors - Get all ancestor terms (all terms higher in the hierarchy)
- get_hpo_parents - Get direct parent terms (one level up)
- get_hpo_children - Get direct child terms (one level down)
- get_hpo_descendants - Get all descendant terms (all terms lower in the hierarchy)
Analysis & Utility Tools
- validate_hpo_id - Validate HPO ID format and verify term exists
- get_hpo_term_path - Get the full hierarchical path from root to a specific term
- compare_hpo_terms - Compare two terms and find their relationship and common ancestors
- get_hpo_term_stats - Get comprehensive statistics and analysis for an HPO term
Prerequisites
- Node.js: Version 18 or higher
- Internet connection: Required for accessing the HPO API
Installation
- Clone or download this server:
cd hpo-server
npm install
- Build the server:
npm run build
Usage
Running the Server
npm start
# or
node build/index.js
Configuration
Add the server to your MCP settings file:
{
"mcpServers": {
"hpo-server": {
"command": "node",
"args": ["/path/to/hpo-server/build/index.js"]
}
}
}
Usage Examples
Search and Discovery
-
Search for seizure-related terms:
Search for HPO terms related to "seizure" using search_hpo_terms -
Search with pagination:
Search for "heart defect" terms with 20 results per page using search_hpo_terms -
Get detailed term information:
Get detailed information about seizure (HP:0001250) using get_hpo_term
Hierarchical Navigation
-
Explore term hierarchy:
Get all ancestors of seizure term using get_hpo_ancestors -
Find related terms:
Get direct children of nervous system abnormality using get_hpo_children -
Browse term descendants:
Get all descendant terms of seizure using get_hpo_descendants
Analysis and Utilities
-
Validate HPO IDs:
Validate if "HP:0001250" is a correct HPO ID using validate_hpo_id -
Get hierarchical path:
Show the complete path from root to seizure term using get_hpo_term_path -
Compare two terms:
Compare seizure and nervous system abnormality terms using compare_hpo_terms -
Get term statistics:
Get comprehensive statistics for seizure term using get_hpo_term_stats -
Batch processing:
Get information for multiple terms (HP:0001250, HP:0000707, HP:0001626) using batch_get_hpo_terms -
List all terms:
Browse all HPO terms with pagination using get_all_hpo_terms
API Features
HPO ID Format Support
HPO IDs can be provided in multiple formats:
- Full format:
HP:0001250 - Short format:
0001250 - The server automatically handles format conversion
Pagination Support
Search and listing tools support pagination:
max- Maximum results per page (varies by tool)offset- Number of results to skip (default: 0)
Parameter Details
- id (string, required for most tools): HPO term ID
- query (string, required for search): Search query - keyword, HPO ID, or synonym
- category (array, optional): Filter by specific HPO categories
- max (number, optional): Maximum results to return
- offset (number, optional): Pagination offset
API Details
- Base URL: https://ontology.jax.org/api/hp/
- Authentication: None required (public API)
- Rate Limiting: Managed by client with 30-second timeout
- Error Handling: Comprehensive error messages and graceful degradation
Project Structure
hpo-server/
├── src/
│ ├── index.ts # Main MCP server implementation
│ ├── types/
│ │ └── hpo.ts # TypeScript interfaces for HPO data
│ ├── utils/
│ │ └── api-client.ts # HPO API HTTP client
│ └── handlers/
│ └── hpo-handlers.ts # MCP tool implementations
├── build/ # Compiled JavaScript output
├── package.json
├── tsconfig.json
└── README.md
Development
Building the Server
cd hpo-server
npm run build
Dependencies
@modelcontextprotocol/sdk- MCP SDK for server implementationaxios- HTTP client for HPO API requests
About the Human Phenotype Ontology
The HPO provides a standardized vocabulary of phenotypic abnormalities encountered in human disease. Each term describes a phenotypic abnormality, such as "Atrial septal defect" or "Intellectual disability". The HPO is:
- A flagship product of the Monarch Initiative
- Part of the Global Alliance for Genomics and Health (GA4GH)
- Actively developed using medical literature, Orphanet, DECIPHER, and OMIM
- Used for phenotype-driven differential diagnostics and genomic analysis
Error Handling
The server includes comprehensive error handling:
- Invalid HPO IDs are properly formatted when possible
- Network errors provide clear user-friendly messages
- API errors are passed through with context
- Parameter validation with helpful error messages
Limitations
- Maximum result limits are enforced to prevent overwhelming responses
- Network timeouts are set to 30 seconds for stability
- Some HPO API endpoints may have their own rate limiting (handled gracefully)
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。