VibeETL: Polars-powered visual Alteryx alternative built by a quadriplegic data scientist
A former data scientist built a blazing-fast ETL tool in 3 months—fully open source.
After a decade as a data scientist and becoming quadriplegic, cardchase dedicated three months to building VibeETL—a visual data manipulation platform designed to outperform legacy ETL tools. The backend is powered entirely by Polars with Rust-native optimizations, leveraging zero-copy Apache Arrow memory transport for extreme speed. The frontend uses React Flow with a custom zero-dependency BFS Topological Layout algorithm to snap nodes instantly without lag. A unique 'Isolate Process Jailing' system runs Python code nodes in ephemeral subprocesses with a strict 30-second timeout, preventing crashes. The UI features SafeInput component shielding to eliminate typing lag on complex forms.
VibeETL is built for the AI era: its manifest-driven Python backend allows developers—and even AI agents—to instantly create new processing blocks by dropping a generated folder into the codebase. While core ingestion, cleansing, and database blocks are stable, the creator needs community help testing external cloud connectors and Gemini Vision AI integration for image captioning pipelines. The tool is fully open source on GitHub (cardchase/VibeETL) and invites contributions from data engineers and cloud architects.
- Backend uses Polars and Rust with zero-copy Apache Arrow memory transport for high-speed data processing.
- Custom zero-dependency BFS layout algorithm in React Flow eliminates lag from legacy libraries like dagre.
- Python code nodes run in isolated ephemeral subprocesses with a 30-second timeout to prevent server crashes.
Why It Matters
An accessible, community-driven ETL tool that brings Polars-level performance to visual data workflows.