llm-moonshot

LLM plugin for Moonshot AI’s models

LLM plugin for models hosted by Moonshot AI.

Installation

First, install the LLM command-line utility.

Now install this plugin in the same environment as LLM:

llm install llm-moonshot

Configuration

You’ll need an API key from Moonshot. Grab one at platform.moonshot.cn.

Set secret key:

llm keys set moonshot

Enter key: <paste key here>

Usage

List what’s on the menu:

llm models list

You’ll see something like:

Moonshot: moonshot/kimi-latest
Moonshot: moonshot/moonshot-v1-auto
Moonshot: moonshot/moonshot-v1-128k-vision-preview
Moonshot: moonshot/kimi-k2-0711-preview
Moonshot: moonshot/moonshot-v1-128k
Moonshot: moonshot/moonshot-v1-32k-vision-preview
Moonshot: moonshot/moonshot-v1-8k-vision-preview
Moonshot: moonshot/moonshot-v1-8k
Moonshot: moonshot/kimi-thinking-preview
Moonshot: moonshot/moonshot-v1-32k
...

Fire up a chat:

llm chat -m moonshot/kimi-k2-0711-preview

Chatting with  moonshot/kimi-k2-0711-preview
Type 'exit' or 'quit' to exit
Type '!multi' to enter multiple lines, then '!end' to finish
> yo moonie
yo! what’s up, moonie?
>

Need raw completion?

llm -m moonshot/moonshot-v1-8k "Finish this haiku: Neon city rain"

Neon city rain,
Glistening streets, a symphony,
Echoes of the night.

Reasoning Content Support

This plugin now supports reasoning content for Moonshot's thinking models (models with "thinking" in the name). When using thinking models, you'll see the model's reasoning process displayed in real-time before the final response:

llm chat -m moonshot/kimi-k2-thinking

[Reasoning] (shown in cyan dim)

The user is asking me to solve a complex problem. Let me think through this step by step...
First, I need to understand the core requirements...
Then I'll analyze the available options...

[Response] (shown in bold green)

Here's my well-reasoned answer to your question...

Available Thinking Models

• moonshot/kimi-k2-thinking - Latest reasoning model
• moonshot/kimi-thinking-preview - Preview reasoning model

The reasoning content helps you understand:

• Decision-making process - See how the model analyzes problems
• Multi-step reasoning - Follow complex thought chains
• Error detection - Catch logical gaps or misunderstandings early

Aliases

Save your wrists:

llm aliases set kimi moonshot/kimi-latest

Now:

llm -m kimi "write a haiku about the AI chatbot Sidney is misbehaving"

Development

Clone, venv, deps—same dance:

git clone https://github.com/ghostofpokemon/llm-moonshot.git
cd llm-moonshot
python3 -m venv venv
source venv/bin/activate
pip install -e .