After this lesson, you will be able to: Deploy Flowise locally, build a document Q&A agent, and expose it via REST API or an embeddable chat widget.
Flowise is LangFlow's friendlier cousin: Node.js-based, faster to install, with a polished UX and one-click chat widget embed. This lesson deploys Flowise, builds a doc Q&A agent, and embeds it in a real webpage.
1. Make sure Node.js 18+ is installed.
2. Run: `npx flowise start`.
3. Open http://localhost:3000.
4. (Optional) Set FLOWISE_USERNAME and FLOWISE_PASSWORD env vars to add basic auth.
Chatflow, a single LangChain chain (e.g., conversational RAG, summarizer). Agentflow, a multi-step agent with tool use, branching, and conditionals. Pick Chatflow for retrieval Q&A; Agentflow when you need real tool-calling reasoning.
1. Click '+ Add New' → 'Chatflow' → blank canvas.
2. Add 'Conversational Retrieval QA Chain'.
3. Add 'ChatOpenAI' (or 'ChatAnthropic'). Add API key as a credential.
4. Add 'In-Memory Vector Store'. Connect to a 'Recursive Character Text Splitter' and a 'PDF File Loader'.
5. Connect 'OpenAI Embeddings' to the vector store.
6. Upload your PDF in the loader's settings.
7. Click 'Save Chatflow'. Test in the right-panel chat.
1. Click 'API Endpoint'. Flowise shows you a curl-ready endpoint.
2. POST to it with `{ "question": "..." }`.
3. Use this in your own apps, the agent is now a queryable API.
Sign in and purchase access to unlock this lesson.