Flowise is a low-code platform for building AI agents and LLM workflows through a visual, node-based editor. The project has been archived and its hosted service is being sunset; the source remains available for self-hosting.

It offers two building modes. Chatflow builds single-agent systems and chatbots with tool calling and retrieval-augmented generation over document sources including TXT, PDF, DOC, HTML, JSON, CSV, Markdown, and SQL. Agentflow builds multi-agent systems with workflow orchestration distributed across coordinated agents, and supports human-in-the-loop review of agent tasks.

The codebase is a single monorepo with a Node backend that serves the API, a React frontend, a components package holding third-party node integrations, and auto-generated Swagger API documentation. Built flows are exposed through a prediction REST API, TypeScript and Python SDKs, and an embeddable chat widget. Execution traces integrate with Prometheus, OpenTelemetry, and other observability tools.

Features

Visual builder: compose agents and workflows from modular nodes in a browser UI
Chatflow: single-agent chatbots with tool calling and RAG over many document formats
Agentflow: multi-agent orchestration across coordinated agents
Human in the loop: review points for agent tasks inside the feedback loop
API, SDK, and embed: prediction API endpoint, TypeScript and Python SDKs, and an embedded chat widget
Observability: full execution traces with Prometheus and OpenTelemetry support
Deployment options: npm, Docker, Docker Compose, and guides for AWS, Azure, GCP, Digital Ocean, Railway, Render, Hugging Face Spaces, and others
Configuration: instance behaviour controlled through environment variables

Claim Flowise

Verify write access with GitHub