Use Kilo Code with TurboLLM
Kilo Code is an agentic coding CLI. One command wires it to your local TurboLLM daemon and launches it — so it runs against whatever model you've loaded, fully offline.
Get started (one command)
From any terminal, with TurboLLM installed:
TurboLLM wires Kilo Code to your local daemon the way the tool expects, then launches it. If it finds an existing config it can't parse, it refuses to overwrite it and prints the manual steps instead — it will never clobber a config it doesn't understand.
What it can do
Everything Kilo Code normally does — reading and editing files, running commands, multi-step agentic tasks — except the model runs on your own GPU. Nothing about your code, prompts, or keys leaves the machine.
Recommended models
Coding agents lean hard on tool-calling and long context. See the Models catalog for the coder picks that pair well with Kilo Code, sized by GPU.
Agentic sessions fill the context window fast. Load the model with a generous context length — TurboLLM's auto-tune shows a VRAM-fit verdict so you can see how much context fits before you commit.
Manual setup (no magic)
The launcher just points Kilo Code at TurboLLM's OpenAI-compatible endpoint — there's nothing it does you can't do by hand. In Kilo Code's provider settings, choose an OpenAI-compatible provider and set:
- Base URL:
http://localhost:6996/v1 - API key: a key from Developer → API Keys
- Model: the model you loaded in TurboLLM
The API key is required only if you've turned on Require API key under Settings → Network (e.g. for LAN sharing); otherwise any non-empty value works.