onix ide

Getting started

After the server is built, the IDE starts it whenever the application opens. The packaged IDE connects to onix-ai-server; the commands below are for a development checkout.

1

Build the server — once is enough

Use the first line for CPU-only operation; use the second for an NVIDIA GPU.

# CPU-only
cd /path/to/onix-ai-core && make server

# or GPU (CUDA)
cd /path/to/onix-ai-core && make server-cuda
2

Verify the connection

Look for the green dot and "Connected to LLangServer" in the lower-left status bar. If it is missing, use AI › Reconnect to LLangServer.

Automatic The IDE starts the server only when it is not already running. It leaves a server started from another terminal or systemd untouched.
3

Load a model

AI › Load Model... — select a .gguf file. No model is loaded by default.

✓

Chat or request a completion

Chat panel: Ctrl+Shift+C — the model responds using its chat template.

Code completion: Ctrl+Space in the editor — generates a raw continuation without chat formatting.

Open this page again from Help › Getting Started