Agentset
The open-source platform to build AI apps that deliver reliable answers.·agentset.ai
Agentset is an open-source platform for building production-ready Retrieval-Augmented Generation (RAG) applications. It connects large language models to a company's own documents, handling ingestion, chunking, embedding, hybrid search, reranking, and automatic source citations. It offers JavaScript and Python SDKs, an MCP server, and both managed cloud and self-hosted/on-premise deployment options.
What it's for
Features 18
- AES-256 encryption and TLS transmission
- Agentic search with tool-calling loop
- AI SDK integration with major model providers
- Automatic source citations
- Bring-your-own-cloud and on-premise deployment
- Document chunk viewer in dashboard
- EU data residency options
- Hosted chat interface with custom domain support
- Hybrid search and reranking
- Ingestion of 22+ document formats
- JavaScript and Python SDKs
- MCP server for connecting knowledge base to external apps
- Metadata filtering
- Model-agnostic: choose your own vector database, embedding model, and LLM
- Multimodal support for images, graphs, and tables
- Preview links for external feedback collection
- Web crawling and ingestion with configurable depth and filters
- YouTube video transcript ingestion
At a glance
Integrations
Complianceself-reported
Pricing
Enterprise plan is custom-priced and available on request.
Forever free tier for personal use and small projects.
1,000 pages, 10,000 retrievals
For production applications ready to scale.
10,000 pages included, unlimited retrievals
Custom pricing, contact sales.
Unlimited pages and retrievals
Special pricing for qualifying startups.