qwen-ai-provider

Qwen AI Provider for interacting with Qwen models via Alibaba Cloud's Model Studio API.

Requirements

This provider requires a valid Qwen API key and adherence to the API specifications provided by Alibaba Cloud.

Installation

The qwen-ai-provider is available as an npm package. You can install it with:

npm i qwen-ai-provider

Provider Instance

You can import the default provider instance qwen from qwen-ai-provider:

import { qwen } from "qwen-ai-provider"

If you need a customized setup, you can import createQwen from qwen-ai-provider and create a provider instance with your settings:

import { createQwen } from "qwen-ai-provider"

const qwenProvider = createQwen({
  // custom settings, e.g.:
  apiKey: "YOUR_QWEN_API_KEY",
  baseURL: "https://dashscope-intl.aliyuncs.com/compatible-mode/v1",
  headers: { "Custom-Header": "value" },
})

The following optional settings can be provided:

• baseURL string

  Use a different URL prefix for API calls, e.g. to use proxy servers. The default is https://dashscope-intl.aliyuncs.com/compatible-mode/v1.

• apiKey string

  Qwen API key, defaults to the DASHSCOPE_API_KEY environment variable.

• headers Record<string,string>

  Custom headers to include in requests.

• queryParams Record<string,string>

  Optional URL query parameters for API calls.

• fetch FetchFunction

  Custom fetch implementation, useful for testing or middleware.

Models

The provider supports text generation/chat and embeddings. The first argument is the model id. For example:

const chatModel = qwen("qwen-plus")

Other model functions include:

• qwen.chatModel(modelId, settings)
• qwen.completion(modelId, settings)
• qwen.textEmbeddingModel(modelId, settings)
• qwen.languageModel(modelId, settings) (alias for chatModel)

Tested Models and Capabilities

This provider has been tested with Qwen models and supports a range of features:

Feature	Support
Text generation	✅
Streaming output	✅ (if supported)
Object generation	✅
Embedding generation	✅

Usage Details

• Chat Settings

  The settings object allows you to fine-tune model behavior, including streaming, temperature control, and tool integration. See inline documentation in the source code for more details.

• Tool Integration

  Some Qwen models support tool usage for function calls. Configure the tools property in your settings as needed.

• Versioning and Packaging

  When publishing packages (e.g. on GitHub Packages), the provider supports pre-release versioning and packaging via GitHub Actions.

Testing

This package includes a suite of tests using Vitest. To run tests:

npm run test

For more details, review the test files in the repository.