Build
Open the canvas, add nodes, assign Python files, and connect them. Shift+drag to create links; use the sidebar to configure each node's properties.
Flowcraft
Visual Flowchart Builder And Executor For Python
Compose, run, and review multi-step Python workflows as visual flowcharts. Wire existing .py scripts together on a canvas, pass data between steps, and inspect every run from a single local interface — no glue code required.
Your Scripts Stay As Scripts
Each node points to a Python file under nodes/. The canvas is the orchestration layer — links define execution order and how return values flow downstream.
Analyze
Flowcraft parses your Python files with AST analysis to infer function signatures, parameters, return values, and input() calls — powering automatic argument wiring.
Run
Execute an entire flow or a single node. Live stdout and stderr stream to the UI. Stop mid-flow, resume from where you left off, and configure timeouts from settings.
Review
Save runs to persist them. Open the dashboard for success rates, timings, and recent failures. Drill into executions to see per-node output, variables, and line traces.
Local Orchestration Without The Overhead
Python file, data save, IF splitter, database interact, API, and AI nodes
Variable resolver hydrates downstream function arguments automatically
SQLite integration with a /database table browser
Scripts explorer to browse, create, and open files in your editor
Transfer page to export and import flowcharts, nodes, and history as ZIP
Get Started In Two Commands
Flowcraft runs locally on your machine. Open the URL shown in the terminal after launch — typically http://localhost:5000.
pip install git+https://github.com/oscarjbyles/flowcraft.git
flowcraftPrototype Pipelines Without Airflow
Built for Python developers, analysts, and solo builders who want visual debugging, run history, and a scripts explorer in one local tool.
Explore The Repository