# Langflow

> Visual builder for AI agents and workflows, with built-in API and MCP servers

- Category: App UIs & Builders / Low-Code & No-Code Builders
- Language: Python
- Licence: MIT
- Repository: https://github.com/langflow-ai/langflow
- Website: https://www.langflow.org
- Canonical page: https://ai-tldr.dev/tools/langflow/

## In short

Langflow is a platform for building and deploying AI-powered agents and workflows. It sits in the Low-Code & No-Code Builders category of the App UIs & Builders tools and is open-source, written in Python and licensed MIT. To get started: install the package.

## Overview

Langflow is a platform for building and deploying AI-powered agents and workflows. It gives developers a visual authoring canvas to assemble flows, plus built-in API and MCP servers that turn every workflow into a tool you can call from an app built on any framework or stack.

It comes with batteries included: support for all major LLMs, vector databases, and a growing library of AI tools. You can drag components together in the visual builder, drop into the underlying Python source to customize any component, and test everything in an interactive playground before you ship.

## What it does

- Visual builder interface to quickly start and iterate on flows
- Source code access so you can customize any component using Python
- Interactive playground to test and refine flows with step-by-step control
- Multi-agent orchestration with conversation management and retrieval
- Deploy a flow as an API, or export it as JSON for Python apps
- Deploy a flow as an MCP server to turn it into a tool for MCP clients

## Getting started

Installing locally is the recommended path. It requires Python 3.10-3.14 and uv (the recommended package manager). If you would rather skip Python setup entirely, Langflow Desktop bundles all dependencies for Windows and macOS.

1. Install the package

   From a fresh directory, install the latest Langflow package with uv.

```shell
uv pip install langflow -U
```

2. Run Langflow

   Start the server. Langflow opens at http://127.0.0.1:7860, ready for you to build.

```shell
uv run langflow run
```

3. Or run with Docker

   Prefer a container? Start Langflow with default settings and open it at http://localhost:7860/.

```shell
docker run -p 7860:7860 langflowai/langflow:latest
```

## When to use it

- Prototype and iterate on multi-agent applications on a visual canvas, then drop into Python to fine-tune individual components
- Ship a finished flow as an API endpoint that any framework or stack can call
- Expose a workflow as an MCP server so MCP clients can use it as a tool
- Build retrieval and conversation-driven assistants that connect to major LLMs and vector databases