01
Install Nock and add Nock Coder
pipx install nock-cli
nock add coder
Needs Python 3.9 or newer. Downloads about 1 GB once.
nock serve gives your tools one local, OpenAI-compatible address. Nothing leaves your machine. About five minutes.
01
pipx install nock-cli
nock add coder
Needs Python 3.9 or newer. Downloads about 1 GB once.
02
nock serve
Keep this terminal open. The address is http://127.0.0.1:8080/v1. It listens on this machine only, refuses requests from websites, and logs nothing you ask.
03
Install the Continue extension, open its config (~/.continue/config.yaml) and add this under models:
- name: Nock Coder
provider: openai
model: nock-coder
apiBase: http://127.0.0.1:8080/v1
apiKey: local
roles: [chat, edit]
Pick Nock Coder in Continue's chat and ask something. The key can be any text; nothing checks it.
04
Anything that lets you set an OpenAI base URL can use the same three values: base URL http://127.0.0.1:8080/v1, any API key, model nock-coder.
from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:8080/v1", api_key="local")
reply = client.chat.completions.create(
model="nock-coder",
messages=[{"role": "user", "content": "Reverse a string in Python."}],
)
print(reply.choices[0].message.content)
It is a 1.5B chat model: it answers questions and writes code, but it cannot drive agent or tool modes. Review code before you run it. Slow answers? Run nock doctor to see which device the engine uses.