1.
RAG (Retrieve and Generate):
The default tool for all assistants, enabling them to retrieve information from a knowledge source (e.g., a Corpus) and generate contextually accurate responses. RAG serves as the foundational tool for most knowledge-based tasks.
2.
Web Search:
This tool enables assistants to fetch real-time information from the web. Users can configure this tool to search through specific engines such as DuckDuckGo or Bing, providing flexibility in how assistants source external information.
3.
API Tool:
The API Tool allows assistants to interact with external APIs to fetch data or perform operations. It is highly configurable, supporting options like:Selecting predefined API types (e.g., Cerner Tools) or custom APIs.
Providing a system prompt to guide interactions.
Specifying authorization details for secure API calls.
This tool is ideal for creating assistants that can interact with enterprise systems or perform domain-specific tasks.
4.
Attachment Tool:
This tool supports indexing documents (e.g., PDFs, Word files) within chat threads and enables assistants to perform interactive chats based on the content of the uploaded documents. It’s particularly useful for handling document-specific queries, like summarizing or explaining key points from attachments.
Assistants can process more advanced workflows.
Specialized tools (like RAG or Web Search) ensure foundational and real-time capabilities.
Custom tools like the API Tool enable integration with external systems, enhancing productivity.
Tools like the Attachment Tool offer document-specific insights, streamlining content-related queries.