Xpert Plugin: Moonshot AI

Overview

@xpert-ai/plugin-moonshot connects Moonshot AI (Kimi) models to the XpertAI platform. The plugin integrates Moonshot's OpenAI-compatible API so XpertAI agents can leverage Kimi's powerful long-context language models with context windows up to 262K tokens, including the latest K2 series with advanced reasoning capabilities.

Core Features

• Ships MoonshotModule, which registers the NestJS provider strategy, lifecycle hooks, and configuration schema required by the plugin runtime.
• Implements MoonshotLargeLanguageModel, a LangChain-powered adapter built on ChatOpenAI that supports streaming chat completions, function calling, and token accounting callbacks for agent telemetry.
• Shares a console-ready moonshot.yaml that drives the XpertAI UI forms (icons, help links, credential prompts) for quick operator onboarding.
• Supports multiple Moonshot model variants including:
  • Moonshot V1 Series: moonshot-v1-8k, moonshot-v1-32k, moonshot-v1-128k for various context length requirements
  • Kimi K2 Series: kimi-k2-0711-preview, kimi-k2-0905-preview, kimi-k2-turbo-preview for enhanced performance
  • Kimi K2.5: kimi-k2.5 with 262K context window, vision support, and deep thinking mode for complex reasoning tasks
  • Kimi K2 Thinking Series: kimi-k2-thinking, kimi-k2-thinking-turbo for advanced reasoning tasks

Installation

npm install @xpert-ai/plugin-moonshot

Peer Dependencies: Ensure your host service also provides @xpert-ai/plugin-sdk, @nestjs/common, @nestjs/config, @metad/contracts, @langchain/openai, chalk, and zod. Refer to package.json for exact versions.

Enabling in XpertAI

1. Declare the plugin before bootstrapping the XpertAI server:

   PLUGINS=@xpert-ai/plugin-moonshot

2. In the XpertAI admin console (or config file), create a model provider pointing to moonshot, then add individual models that map to the specific Moonshot/Kimi versions you want to use.

Credentials & Model Configuration

The moonshot.yaml schema backs the form fields you see in the console:

| Field | Description | | ---------- | -------------------------------------------------------------------------------------------------------------- | | api_key | Required. Your Moonshot API Key from platform.moonshot.cn/console/api-keys. | | base_url | Optional. Base URL for API requests (defaults to https://api.moonshot.cn/v1). Useful for proxy configurations. |

During validation, the plugin instantiates a ChatOpenAI client with your credentials and sends a test message ("你好") to ensure connectivity and permissions.

Model Capabilities

• Long Context Support: All Moonshot models excel at handling long-context scenarios, with the V1-128K variant supporting up to 128,000 tokens and the latest K2.5 supporting up to 262,000 tokens of context.
• Conversational Models: MoonshotLargeLanguageModel merges provider credentials with per-model overrides, enables streaming, and registers token usage callbacks so agent telemetry stays accurate.
• Advanced Reasoning: The K2 Thinking series models provide enhanced reasoning capabilities for complex problem-solving tasks.
• OpenAI Compatibility: Built on LangChain's ChatOpenAI, ensuring seamless integration with existing OpenAI-compatible workflows.

Development & Debugging

From the repo root, run Nx commands for this package:

cd xpertai
npx nx build @xpert-ai/plugin-moonshot
npx nx test @xpert-ai/plugin-moonshot

Artifacts land in xpertai/models/moonshot/dist. Jest settings live in jest.config.ts, and the asset copier runs via npm run prepack before publishing.

License

This plugin is distributed under the AGPL-3.0 License located at the repository root.