Capability · Framework — orchestration

Langflow

Langflow exposes LangChain's primitives as a drag-and-drop canvas. Every node is a real Python class, so flows can be exported as Python code or served as an API. The project has grown into a general AI-app canvas with first-class support for multiple vector DBs, tools, and custom components.

Framework facts

Category: orchestration
Language: Python
License: MIT
Repository: https://github.com/langflow-ai/langflow

Install

pip install langflow
langflow run
# open http://127.0.0.1:7860

Quickstart

# From the UI, build a flow with a ChatInput -> LLM -> ChatOutput
# Then call it as an API:
import requests
requests.post(
  'http://127.0.0.1:7860/api/v1/run/<flow-id>',
  json={'input_value': 'hello world', 'output_type': 'chat'}
).json()

Alternatives

Flowise — TypeScript sibling
Rivet — desktop-native visual IDE
PromptFlow — Microsoft / Azure-integrated
Dify — product-style builder

Frequently asked questions

Does Langflow lock me in?

No. Every flow can be exported as a JSON you can import elsewhere, or as plain LangChain Python code — so you can outgrow the canvas without rewriting.

Langflow vs Flowise?

Langflow is Python/LangChain-py, ideal for teams on a Python data stack. Flowise is TypeScript/LangChain.js, a better fit for Node apps. Feature parity is similar; pick by runtime.

Sources

Langflow — docs — accessed 2026-04-20
Langflow GitHub — accessed 2026-04-20