Langflow

Langflow is an open source, Python-based visual builder for AI agents and LLM applications. Instead of writing orchestration glue code, you design multi-agent workflows and RAG pipelines by arranging components on a drag-and-drop canvas — then deploy with one click, where every flow automatically becomes a REST API endpoint callable from any framework or stack. The project is stewarded by DataStax (the Astra DB company) and offers a free, enterprise-grade managed cloud alongside self-hosting.

Key features:

• Visual builder with an interactive playground for testing and refining flows step by step
• Full source code access — any component can be customized or replaced with Python
• Support for all major LLM providers, vector databases and a growing library of AI tools
• Multi-agent orchestration with conversation management and retrieval
• Deploy flows as a REST API or as an MCP server, so external clients can call your flows as tools
• Observability integrations with LangSmith, Langfuse and others
• Export flows as JSON or Python code to embed in your own applications

Self-hosting options:

• pip / uv — uv pip install langflow then uv run langflow run (requires Python 3.10–3.14); starts at http://127.0.0.1:7860
• Docker — docker run -p 7860:7860 langflowai/langflow:latest
• Langflow Desktop — a native Windows/macOS app with all dependencies bundled, no Python environment needed
• Deployment guides exist for all major clouds, or use DataStax's free managed cloud

Compared with hosted agent-builder SaaS, Langflow trades a locked-down web console for openness: it is MIT-licensed, model-, API- and database-agnostic, and stays extensible with custom Python components. It's a natural landing spot for teams coming from Flowise, which was discontinued in 2026 — the drag-and-drop model is similar, while Langflow remains under active development.