Agentset is an open-source platform designed for building, evaluating, and deploying production-ready Retrieval Augmented Generation (RAG) and agentic applications. It offers end-to-end tooling, catering to developers seeking a comprehensive solution for ingestion, indexing, evaluation, and hosting, with a focus on a strong developer experience and built-in features like citations.

How It Works

The platform provides a turnkey RAG pipeline, handling ingestion, chunking, embeddings, and retrieval. It emphasizes model agnosticism, allowing users to integrate their preferred LLMs, embedding models, and vector databases. Agentset includes tools for quality tracking via benchmarks and evaluations, a chat playground with message editing and citations, and production hosting capabilities with preview links and custom domains. The backend is built with TypeScript, Next.js, AI SDK, Prisma, Supabase, and Trigger.dev.

Quick Start & Requirements

Local development requires pnpm for dependency management. Setup involves copying the environment file (cp .env.example .env), installing dependencies (pnpm install), deploying database migrations (pnpm db:deploy), and starting the application (pnpm dev:web). Useful scripts include pnpm db:studio for database management and pnpm generate-openapi for regenerating the OpenAPI specification. A cloud version with a free tier is available at https://app.agentset.ai/login, and self-hosting documentation is provided at https://docs.agentset.ai/self-hosting.

Highlighted Details

• Supports built-in citations and message editing in its chat playground.
• Handles over 22 file formats for data ingestion.
• Features include partitions, an MCP server, and deep research capabilities.
• Offers production hosting with preview links and custom domains.
• Provides an API with typed SDKs and an OpenAPI specification.
• Includes built-in multi-tenancy support.

Maintenance & Community

The project maintains an active community on Discord for support and discussion. Contributions are welcomed via issues and pull requests. The project encourages users to star the repository to support its development.

Licensing & Compatibility

Agentset is released under the MIT license, permitting broad use and modification, including for commercial purposes.

Limitations & Caveats

No specific limitations or caveats were explicitly detailed in the provided README content.