About

Falang is a universal solution for creating flowcharts

It is built on the basis of web technologies, which allows it to be used both in desktop and web applications.

Installation

The application is available for Windows, Linux, MacOS. To install you need to download distribution and run it.

Download link

.

After installation, run the program and choose File → New Project… (or click New Project… on the welcome page). In the dialog, enter a name, a folder and a project type. An old project made with a previous version of Falang is converted automatically the first time you open it; a backup of the original files is kept in its backup folder.

Project types

The Falang desktop app has three kinds of projects:

• Text
  • allow you to visualize algorithms in human language.
• Logic constructor
  • universal business logic constructor, with the ability to generate code in various programming languages: C++, Rust, Go, C#, TypeScript.
• Code

  (C++, JavaScript, TypeScript, PHP or Rust — chosen when the project is created) - algorithm schemes whose icons hold raw source code in that language, with syntax highlighting. The code is not checked; Project → Export Code Documents assembles one source file per scheme into the project’s generated folder.

Two more products are built on the same scheme engine:

• Falang for Arduino
  • a separate desktop app that compiles schemes into an Arduino sketch, uploads it to the board and debugs it right on the hardware.
• Workflow builder
  • a web editor for automations (integrations between services, bots, AI scenarios) with durable execution on Temporal.

Desktop app features

• Project files. A project is a plain folder: falang.json describes the project, and every scheme is a separate JSON file under falang/schemes/, named after the scheme — folders of the project tree are real folders on disk. It is convenient to keep a project in git and to read it with other tools.

• Version history. Project → Version History (Sketch → Version History in the Arduino app) lists saved versions, compares any two of them side by side and restores an older one. Versions are stored in a git repository; the repository and the author are set in Settings → Versioning….

• AI agent. Project → Agent opens a chat that edits schemes from a plain-language description, the same way as the workflow builder’s agent. It needs your own OpenAI-compatible endpoint, model and API key, set once in Settings → Agent….

• Coding agents. The app puts a .mcp.json and a CLAUDE.md into every project folder (only if they are not there yet). An MCP-capable coding agent such as Claude Code, started in that folder, gets tools to read, edit and validate the schemes. While it edits a scheme, the open tab is locked and shows an overlay.

• Export to PDF
  • print schemes, one page per scheme.

• Interface language - Settings → Language… (English or Russian).

About the author

Project author: Sachik Sergey.

Email for communication: sachik-sergey@yandex.ru.