# Agentset RAG infrastructure for developers building AI chat, search, and document Q&A. - Findings24: - Website: - Pricing: Freemium - Categories: AI Knowledge Base, AI Developer Tools, AI Search API ## What is Agentset? Agentset is an open-source platform for developers building retrieval-augmented generation (RAG) applications. It provides infrastructure for search and question answering across a product's own documents, reducing the need to assemble and maintain a separate ingestion and retrieval pipeline. The platform combines document processing, retrieval, and answer generation for applications with growing document collections and usage. Developers can upload data through JavaScript or Python SDKs and connect Agentset to their applications using the AI SDK integration, or expose a knowledge base through its MCP server. A customizable chat interface with preview links lets teams collect external feedback before building a complete user interface. Agentset offers managed deployment, with on-premise and bring-your-own-cloud options on its Enterprise plan. Its founders describe building the product after spending three months improving a RAG system for nine million pages, with the goal of making that infrastructure available to other engineering teams. ## What can Agentset do? ### Hybrid retrieval and agentic reasoning Combines hybrid search, reranking, and built-in agentic reasoning to retrieve relevant material and support answers over a knowledge base. Semantic search is included across all pricing plans. ### Multimodal document processing Processes images, graphs, and tables alongside text. The site lists support for more than 22 file formats, including PDF, DOCX, XLSX, PPTX, HTML, CSV, and common image formats, so a knowledge base can include varied business documents. ### Automatic source citations Adds source citations to generated answers so users can inspect the material behind a response. Citations and sources are included on Free, Pro, and Enterprise plans. ### Metadata filtering Supports filtering by document metadata to restrict retrieval and answers to a selected subset of the knowledge base. Developers can attach metadata when creating an ingestion job. ### Developer integrations Provides JavaScript and Python SDKs for data uploads, an AI SDK integration for application development, and an MCP server for making a knowledge base available to external applications. ### Configurable AI stack Lets developers choose their vector database, embedding model, and large language model. This supports integration with an existing AI stack and changing individual components as application requirements evolve. ## Use Cases ### Legal research within a product Add document search and question answering to a legal information service. Agentset cites Gyldendal Rettsdata as a customer using the platform for a legal corpus of three million pages. ### Research-grounded medical Q&A Help users inspect the research underlying answers in a medical application. Mederva Health describes using Agentset to make its answers more reliable and grounded in research. ### Municipal content and image search Search municipal document collections where relevant information may appear in images as well as text. SustainBridge describes using Agentset for complex image search across content spanning hundreds of pages. ### Evaluating a document assistant Share a customizable chat preview with external reviewers to gather feedback on how an assistant answers questions over a team's documents before committing to a complete application interface. ## Frequently Asked Questions ### How much do the Free and Pro plans include? Free costs $0 and includes 1,000 pages, 10,000 retrievals per month, and community support. Pro costs $49 per month and includes 10,000 pages, unlimited retrievals, and email support. Additional Pro pages cost $0.01 per page per month. Deep Research is available on Pro and Enterprise, but not Free. ### How does Agentset count a page? Agentset counts every 1,000 characters of parsed text as one page. Usage can therefore differ from the visible page count in a PDF, PowerPoint presentation, or Word document. ### Can Agentset automatically sync external data sources? Connectors automatically sync content from external sources such as Google Drive, SharePoint, or Notion into a knowledge base. They are not included on Free. Pro connectors cost $100 each per month; Enterprise connector pricing is custom. ### What deployment options are available? Managed deployment is included on Free, Pro, and Enterprise. The pricing comparison reserves self-hosted deployment for Enterprise, which also offers on-premise or bring-your-own-cloud deployment, single sign-on, custom integrations, custom document formats, and dedicated engineering support. Enterprise pricing is available through sales. ### Can I change plans or get an organization discount? Plans can be changed at any time. Upgrades provide immediate access to the new features; the pricing FAQ says downgrades receive a prorated bill based on usage. Qualifying startups and nonprofit organizations can request special pricing. ## Quick Facts - **Category**: RAG infrastructure for AI search and Q&A - **Primary users**: Developers and engineering teams - **Source availability**: Open-source - **SDK languages**: JavaScript and Python - **Website**: agentset.ai