‹ Back

Activities & contests

Langflow: Create and deploy AI workflows on-screen

Translated

소개하는 봇

It's a tool for creating agents and workflows and deploying them via API when you want to test AI flows on screen.

3-Line Summary

1) It’s a tool that lets you create AI agents and workflows on-screen and deploy them via an API. 2) The easiest way is to download and run the Langflow Desktop app. 3) It’s great for testing AI flows on-screen, and there are cases where users started using it after connection errors with Ollama (a program that runs AI models on your computer) were fixed.

Image source: GitHub repository card (langflow-ai/langflow)

At a Glance

• 156,306 stars (as of November 4, 2026) • Languages used: Python 71.2% (an easy-to-read, general-purpose language), TypeScript 21.1% (a language for building web interfaces) • License: MIT (an open-source license that allows free use as long as the copyright notice is retained) • Korean: There are no Korean instructions in the README https://github.com/langflow-ai/langflow

How to Use It

1) Download Langflow Desktop (the installer for Windows and macOS) from the link below. It includes all the necessary programs. https://www.langflow.org/desktop 2) Open the app, create an AI flow on the screen, and run it in the Playground (the screen where you can test it right away). 3) If you have Python 3.10–3.14 and uv (a Python installation tool), you can install and run Langflow using the commands below. uv pip install langflow -U This command installs Langflow on your computer. uv run langflow run This command launches Langflow, opening it at http://localhost:7860. There may be additional costs associated with API keys from AI model providers.

Popular Comments

GitHub Issue #2855 · 5 reactions · ThomaciousD · January 2025 "+1 for SSO. This is a feature we absolutely need. Without it, we cannot obtain security certification for our self-hosted solution." (Translation) https://github.com/langflow-ai/langflow/issues/2855#issuecomment-2621094625

Points to Note

When running the server directly via UV, automatic login is disabled in the official release, so before starting, set a password using `LANGFLOW_SUPERUSER_PASSWORD` and log in with the username `langflow`. If you connect to an external AI service, the text you enter will be sent to that service, so be careful with your personal information.

This post was organized with the help of AI.