Getting Started
- Install via pip — Run
uv pip install langflow -U(requires Python 3.10–3.13) or download Langflow Desktop for Windows/macOS - Launch the UI — Start the local server and open the visual builder in your browser
- Build a flow — Drag and drop components (LLMs, vector stores, tools) onto the canvas and connect them
- Deploy — Export your flow as a REST API, JSON, or as an MCP server for use in any application
Key Features
- Visual Builder — Drag-and-drop canvas lets you compose AI workflows without writing boilerplate code
- Latest release (March 2026) — The most recent release brings continued improvements to component stability, MCP server deployment, and provider integrations.
- Source Code Access — Every component can be customized with Python for full flexibility
- Multi-Agent Orchestration — Build and manage multiple agents with conversation history and retrieval support
- API & MCP Deployment — Instantly expose any workflow as a REST API endpoint or Model Context Protocol server
- Broad Integrations — Supports all major LLMs, vector databases, and observability tools like LangSmith and LangFuse
- Interactive Playground — Test and debug flows step-by-step in real time before deploying
- Latest release (March 2026) — The most recent release brings continued improvements to component reliability, MCP server support, and the visual editor experience.
// related tools
AutoGPT
AI / Agents & Automation
Open-source platform to build, deploy, and run autonomous AI agents
oss
web git
CrewAI
AI / Agents & Automation
Framework for orchestrating autonomous AI agent teams
oss
web git
Dify
AI / Agents & Automation
Open-source platform for building LLM apps with visual workflows
oss
web git