Runify
Lightweight White-box AI Agent Platform
Low footprint · Visual building · Fully controllable custom AI Agents
💡 No more black boxes and magic — everything under your control. In Runify, every reasoning step, every tool call and every decision node is fully transparent, editable and debuggable. We hand the determinism of algorithms and engineering back to developers.
What problem it solves
Most Agent frameworks hide reasoning inside a black box, making them hard to debug and control. Runify lays every decision node out on a visual canvas, letting you compose Agents like building blocks and step in, edit and debug at any time.
Feature overview
| Feature | Description |
|---|---|
| White-box | Reasoning, tool calls and decision nodes are transparent, editable, debuggable |
| Lightweight | 100MB–200MB memory, 1–2s startup; a 1-core 2GB box handles local streaming |
| Visual building | Drag-and-drop DAG canvas — no complex BPMN |
| Smart notes | Built-in rich-text editor as Agent context and personal knowledge base |
| Data sources | One-click connect to PostgreSQL / MySQL / Redis and more |
| Cross-platform | Responsive web + Electron desktop (macOS / Windows) |
Next steps
- Set up your environment in Installation
- Run your first Agent in Quick Start
- Go deeper with the Workflow Guide
This documentation targets v0.1.0-beta, an early release whose API and features may still change. Use the version dropdown in the top-right to switch versions.