教程导航

Python API: API参考 | MCP setup: MCP Support

核心概念

• Loading Tools → 工具加载教程 - Complete tutorial to loading tools with Python API and MCP terminal commands

• Listing Tools → 工具清单教程 - Discover and filter tools by capability, domain, and IO

• Tool Caller → 工具调用教程 - Primary execution engine with dynamic loading, validation, and MCP server integration

• Tool Composition → 工具组成教程 - Chain ToolUniverse’s 1000+ tools into powerful scientific workflows using Tool Composer

• Coding API → 编码 API - 类型化函数 - Import and call tools like normal Python functions with type safety

• Scientific Workflows → 科学工作流 - Real-world research scenarios: drug discovery, safety analysis, literature review

• MCP Support → MCP Support - Model Context Protocol integration and server setup

• MCP Name Shortening → MCP Tool Name Shortening - Automatic tool name shortening for MCP 64-character limit compatibility

• MCPB Support → MCP Bundle - Standalone executable bundle for Claude Desktop and other clients

• HTTP API → HTTP API - Remote Access - Remote access via HTTP/REST with auto-updating server and minimal client dependencies

• Compact Mode → Compact Mode Guide - Optimize context window usage by exposing only core tools (4 tools) while maintaining full functionality

• Streaming Tools → 流媒体工具教程 - Real-time streaming output and custom tool integration

• Logging → <h1>日志记录教程</h1> - Comprehensive logging configuration and debugging

• Result Caching → 结果缓存 - Configure in-memory and persistent caches for tool results

• Interaction Protocol → AI工具交互协议 - Understanding tool interaction patterns

工具发现与使用

• Tool Discovery → Tool Finder Tutorial - Tutorial to ToolUniverse’s three tool finder methods: keyword, LLM, and embedding search

• Tools Overview → 可用工具参考 - Comprehensive overview of all available tools

AI Agent Platform Setup

• Platform Setup → Set Up ToolUniverse - Connect ToolUniverse to your AI agent

LLM Providers

• vLLM Support → vLLM Support - Use self-hosted LLM models with vLLM for high-performance inference

• OpenRouter Support → OpenRouter 支持 - Access multiple LLM providers through OpenRouter API

高级功能

• Hooks System → 后处理工具输出 - Intelligent output processing with AI-powered hooks

  • SummarizationHook → 摘要钩子 - AI-powered output summarization

  • FileSaveHook → 文件保存钩子 - File-based output processing and archiving

  • Hook Configuration → 钩子配置 - Advanced configuration and customization

  • Server & Stdio Hooks → 服务器与标准输入输出钩子集成 - Using hooks with server and stdio interfaces

工具集合

专为特定研究领域设计的专业工具集合：

• Clinical Guidelines → 临床指南工具 - Search and extract clinical practice guidelines from NICE, WHO, PubMed, and 5 other authoritative sources

• Literature Search → Literature Search Tools Tutorial - Comprehensive literature search across PubMed, arXiv, bioRxiv, and academic databases

• Profile & Workspace → Profile & Workspace - Configure which tools to load, cache settings, LLM providers, and hooks via profile.yaml. Includes life-science.yaml, a ready-made profile covering all life science and general tools.

• tu CLI → tu — ToolUniverse CLI - Discover and execute tools from the shell without an MCP server.