Agentset

Description

Agentset is an open-source platform for building AI apps that deliver reliable answers over a knowledge base. It packages production-grade retrieval-augmented generation (RAG) so developers can go from prototype to production accuracy without deep RAG expertise, using advanced retrieval, high-resolution parsing, ranking, and an agentic mode that plans and reasons across multiple steps.

The platform natively handles multimodal content, working with images, graphs, and tables alongside text, and automatically cites the sources of its answers so users can inspect them. It supports metadata filtering to base answers on a subset of data, customizable preview links for gathering feedback, and a hosted, shareable chat and search interface.

Agentset is built for developers, with JavaScript and Python SDKs, support for 22+ file formats, an MCP server to bring a knowledge base into external applications, and AI SDK integration. It is model-agnostic, letting teams choose their own vector database, embedding model, and LLM, and offers flexible deployment on Agentset Cloud, bring-your-own-infrastructure, or on-premise.

Pricing scales from a forever-free plan with 1,000 pages and 10,000 retrievals, to a Pro plan for production applications with 10,000 pages and unlimited retrievals, up to an Enterprise plan with unlimited usage, on-premise or BYOC deployment, and SOC 2, HIPAA, and GDPR reports.

Agentset's Core Features

• Production-grade RAG with advanced retrieval and ranking

• Agentic mode that plans, reasons, and evaluates across steps

• Native multimodal support for images, graphs, and tables

• Automatic source citations for every answer

• Metadata filtering to scope answers to a data subset

• JavaScript and Python SDKs with 22+ file formats

• MCP server and AI SDK integration

• Model-agnostic choice of vector DB, embedding model, and LLM

How to use Agentset?

1. 1

   Sign up: Start on the forever-free plan or schedule a demo.

2. 2

   Upload your data: Ingest documents in any of 22+ supported file formats.

3. 3

   Configure retrieval: Choose your vector database, embedding model, and LLM.

4. 4

   Build the experience: Use the SDKs, AI SDK, or MCP server, or enable the hosted chat and search UI.

5. 5

   Ship and share: Deploy on Agentset Cloud, your own infrastructure, or on-premise, and share preview links.

Agentset's Use Cases

• AI knowledge assistant
• Document search
• Legal and research corpora
• Customer-facing chat
• Agentic RAG apps

FAQ from Agentset

• What is Agentset?

  Agentset is an open-source, RAG-as-a-service platform that helps developers build AI chat and search apps that deliver reliable, cited answers over their own data without RAG expertise.

• Does it support more than plain text?

  Yes. Agentset natively works with images, graphs, and tables just like text, so answers can draw from every part of a knowledge base, and it supports 22+ file formats.

• Can I use my own models and databases?

  Yes. Agentset is model-agnostic, letting you select your own vector database, embedding model, and LLM.

• How can I integrate it into my app?

  You can use the JavaScript and Python SDKs, integrate with the AI SDK, or connect your knowledge base to external applications through the MCP server.

• What is a page for pricing purposes?

  Agentset counts every 1,000 characters of parsed text as a page, which may differ from the number of pages in your source PDF, PowerPoint, or Word document.

• How is Agentset priced?

  There is a forever-free plan with 1,000 pages and 10,000 retrievals, a Pro plan with 10,000 pages and unlimited retrievals, and an Enterprise plan with unlimited usage and on-premise or BYOC deployment.