A visual builder for LLMs
In April 2023, Henry Heng published Flowise, an open source visual builder for applications based on LangChain.js. The project was born to bridge the gap between those familiar with LangChain primitives (chain, agent, retriever) and those who want to compose them visually without writing TypeScript code. In 2024, Flowise became a commercial offering through FlowiseAI Inc., which keeps the community edition open and develops enterprise features in parallel.
Architecture and stack
Flowise is written in TypeScript on the Node.js runtime. The interface is a drag-and-drop canvas where each node represents a LangChain.js primitive: LLM provider, prompt template, memory, retriever, chain, agent, tool. By connecting nodes, the application’s execution graph is built, which Flowise runs on each request.
Every flow can be exposed as an HTTP API or as an embeddable chat widget for websites. The engine supports all models integrated in LangChain.js (OpenAI, Anthropic, Hugging Face, Azure, Cohere, Ollama and others) and most available vector databases.
Licence and positioning
The community edition is distributed under the Apache 2.0 licence, with full rights to modify, redistribute and use commercially. The distinction with the enterprise version concerns operational features (SSO, audit log, organisation management) but not the core of the engine.
Marketplace and templates
Flowise includes a marketplace of public templates — customer support chatbots, Q&A over documents, agents with tool calling, RAG pipelines — that allow starting from a working configuration and adapting it to the specific use case. In the LangChain visual builder segment, Flowise is a direct competitor to LangFlow: the choice between the two depends mainly on the preferred stack (TypeScript/Node.js for Flowise, Python for LangFlow) and on specific required integrations.
Link: flowiseai.com