What is Agentset?
Agentset is a platform for developers building AI chat and search applications over their own data. It provides retrieval-augmented generation, document processing, search and answer delivery with citations, without requiring teams to assemble the full RAG stack themselves. Teams can upload content through JavaScript or Python SDKs, choose their vector database, embedding model and language model, and add a customizable chat interface or connect the knowledge base to external applications through an MCP server.
Agentset Features:
Hybrid Search: Agentset combines hybrid search and reranking to retrieve relevant material from a knowledge base for AI-generated answers.
Multimodal Support: The platform can work with images, graphs and tables as well as text, allowing questions to draw on more of a document collection.
Cited Answers: Answers include citations to their sources so users can inspect the material behind a response.
Metadata Filters: Metadata filtering lets teams limit answers to a selected subset of the available data.
File Ingestion: JavaScript and Python SDKs support uploading data in more than 22 file formats.
Model Choice: Teams can select their own vector database, embedding model and large language model for an application.
External Connections: An MCP server connects a knowledge base to external applications, while an AI SDK integration supports use in custom applications.
Agentset Benefits:
Less Setup: Developers can start with an existing RAG system instead of building and maintaining every retrieval component themselves.
More Trust: Source citations give people a way to check an answer against its underlying material before relying on it.
Broader Coverage: Support for visual and structured content can help teams answer questions that text-only search may miss.
Focused Results: Filtering makes it possible to keep answers within the relevant part of a knowledge base.
Stack Flexibility: Choosing models and a vector database lets teams adapt the system to their existing technical preferences.
Use Cases:
Legal Search: Build a search experience over a large legal corpus so users can ask questions and check cited source material.
Medical Research: Create a question-answering tool that grounds responses in research documents for teams working with medical information.
Municipal Content: Help residents or staff find information across municipal documents, including content that uses images.
Product Knowledge: Add AI chat and search to an application that answers user questions from its own documentation or other uploaded content.
Custom AI Apps: Use the SDKs and model choices to build a tailored retrieval-based application around a business knowledge base.