K-Code
The missing link between your code and the hardware. Develop, debug, and deploy Kneron AI applications without ever leaving Visual Studio Code.
Development Stack

The Context Switch Tax
Developing for Edge AI usually involves a messy workflow:
- Write code in IDE.
- Switch to Terminal to run Docker commands.
- Wait for compilation (blindly).
- Manually copy binary to device via SCP/ADB.
- Read logs in a separate serial console.
This friction breaks your "flow state" and slows down iteration cycles.
The Integrated Workflow
K-Code unifies the entire toolchain into the VS Code Side Bar.
- Zero-Config Environment: K-Code manages the Docker containers for you. No more complex
docker run -v ...commands. - One-Click Build: Press
F5to compile the model, flash the firmware, and start the debugger in one action. - IntelliSense: Autocompletion for `kneron_api` makes learning the SDK trivial.
Features for Production
Built for the "AI Application Developer".
Project Scaffolding
Start in seconds.
Don't copy-paste folder structures. Use the "New Kneron Project" wizard to generate a production-ready template with:
- • Valid `requirements.txt`
- • Optimized Dockerfile
- • Sample YOLO/ResNet inference scripts
Device Manager
Goodbye, `lsusb`.
The sidebar lists all connected Kneron accelerators. See their status (Idle/Busy), Firmware Version, and Serial Number instantly.
Action: Right-click a device to "Update Firmware" via DFU.
Transparent Container
We handle the complexity.
K-Code silently mounts your workspace into the Kneron Toolchain container. You get the isolation and reproducibility of Docker with the ease of use of a local tool.
Supports seamless cross-compilation for KL520, KL720, and KL530.
Intelligent Coding
Write code faster with context-aware snippets and type checking.
Hover over any `kneron_api` function to see inline documentation, parameter types, and return values without opening the browser.
Try the IDE Experience
See how K-Code integrates seamless device management into VS Code.
Launch IDE Mockup