mcp-pubmed-server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HTTP_PROXY | No | HTTP代理URL | |
| HTTPS_PROXY | No | HTTPS代理URL | |
| PUBMED_EMAIL | Yes | 邮箱地址 | |
| ABSTRACT_MODE | No | 摘要截断模式：quick（1500字符）| deep（6000字符） | quick |
| FULLTEXT_MODE | No | 全文模式：disabled | enabled | auto | disabled |
| PROXY_ENABLED | No | 代理开关：enabled | disabled | disabled |
| PROXY_TIMEOUT | No | 代理超时设置（秒） | 30 |
| ENDNOTE_EXPORT | No | EndNote导出：enabled | disabled | enabled |
| PROXY_PASSWORD | No | 代理密码 | |
| PROXY_USERNAME | No | 代理用户名 | |
| PUBMED_API_KEY | Yes | NCBI API密钥 | |
| PROXY_RETRY_COUNT | No | 代理重试次数 | 3 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| pubmed_searchC | 搜索PubMed文献并返回结构化数据，供LLM进一步分析 |
| pubmed_quick_searchB | 快速搜索PubMed文献，返回精简结果，优化响应速度 |
| pubmed_cache_infoA | 获取缓存统计信息和状态，支持内存和文件缓存管理 |
| pubmed_get_detailsA | 获取指定PMID的完整文献信息，包括全文摘要和详细元数据 |
| pubmed_extract_key_infoC | 提取文献关键信息，优化LLM理解和处理 |
| pubmed_cross_referenceC | 交叉引用相关文献，用于事实核查和深度分析 |
| pubmed_batch_queryB | 批量查询多个PMID的详细信息，优化上下文窗口使用 |
| pubmed_detect_fulltextC | 检测文献的开放获取状态和全文可用性 |
| pubmed_download_fulltextB | 下载指定文献的全文PDF（如果可用） |
| pubmed_fulltext_statusC | 获取全文缓存状态和统计信息 |
| pubmed_batch_downloadB | 批量下载多个文献的全文PDF，支持跨平台智能下载 |
| pubmed_system_checkA | 检查系统环境和下载工具可用性 |
| pubmed_endnote_statusC | 获取EndNote导出状态和统计信息 |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 13 tools
Most tools have clear boundaries (e.g., get_details vs batch_query), but pubmed_quick_search and pubmed_search both perform searches with different output formats, which could cause misselection. The descriptions help differentiate them, so the ambiguity is moderate.
All tools share the pubmed_ prefix, which is a strong convention, but the internal structure mixes verb-led (search, get_details, extract_key_info) and noun-led (cache_info, fulltext_status, endnote_status) styles. This inconsistency makes the naming less predictable than a uniform verb_noun pattern.
With 13 tools, the server is within the ideal 3-15 range and covers search, details, batch, and download workflows. However, four status/check tools (cache_info, fulltext_status, endnote_status, system_check) feel slightly redundant, making the count a bit heavier than necessary.
The surface covers the main PubMed workflows: searching, fetching details, batch queries, full-text detection and download. The main gap is an EndNote export tool—only the status is exposed—and there is no tool for citation export or advanced query building, but these are minor for the core purpose.