Zhipu AI Provider - Vercel AI SDK Community Provider

This is a Zhipu prodiver for the AI SDK. It enables seamless integration with GLM and Embedding Models provided on bigmodel.cn.

Setup

# npm
npm i zhipu-ai-provider

# pnpm
pnpm add zhipu-ai-provider

# yarn
yarn add zhipu-ai-provider

Set up your .env file / environment with your API key.

ZHIPU_API_KEY=<your-api-key>

Provider Instance

You can import the default provider instance zhipu from zhipu-ai-provider (This automatically reads the API key from the environment variable ZHIPU_API_KEY):

import { zhipu } from 'zhipu-ai-provider'

Alternatively, you can create a provider instance with custom configuration with createZhipu:

import { createZhipu } from 'zhipu-ai-provider';

const zhipu = createZhipu({
  baseURL: "https://open.bigmodel.cn/api/paas/v4",
  apiKey: "your-api-key"
});

You can use the following optional settings to customize the Zhipu provider instance:

• baseURL: string
  • Use a different URL prefix for API calls, e.g. to use proxy servers. The default prefix is https://open.bigmodel.cn/api/paas/v4.
• apiKey: string
  • Your API key for Zhipu BigModel Platform. If not provided, the provider will attempt to read the API key from the environment variable ZHIPU_API_KEY.
• headers: Record<string,string>
  • Custom headers to include in the requests.

Language Model Example

import { generateText } from 'ai';
import { zhipu } from 'zhipu-ai-provider';

const { text } = await generateText({
  model: zhipu('glm-4-plus'),
  prompt: 'Why is the sky blue?',
});

console.log(result)

Embedding Example

const { embedding } = await embed({
  model: zhipu.textEmbeddingModel("embedding-3", {
    dimensions: 256, // Optional, defaults to 2048
  }),
  value: "Hello, world!",
});

console.log(embedding);

Image Generation Example

Zhipu supports image generation with the cogview models, but the api does not return images in base64 or buffer format, so the image urls are returned in the providerMetadata field.

import { experimental_generateImage as generateImage } from 'ai';
import { zhipu } from 'zhipu-ai-provider';

const { image, providerMetadata } = await generateImage({
  model: zhipu.ImageModel('cogview-4-250304'),
  prompt: 'A beautiful landscape with mountains and a river',
  size: '1024x1024',  // optional
  providerOptions: {  // optional
      zhipu: {
          quality: 'hd'
      }
  }
});

console.log(providerMetadata.zhipu.images[0].url)

Features Support

• Text generation
• Text embedding
• Image generation
• Chat
• Tools
• Streaming
• Structured output
• Reasoning
• Vision
• Vision Reasoning
• Provider-defined tools
• Voice Models

Documentation

• Zhipu documentation
• Vercel AI SDK documentation
• Zhipu AI Provider Repo