Z.ai (GLM)

LLM support via Z.ai's GLM family. GLM 5.2 is an open-weight (MIT-licensed) mixture-of-experts model with a 1M-token context window and strong tool calling at a fraction of frontier-model cost — a compelling choice for cost-sensitive voice agents that still need reliable function calling.

Official Documentation

• Z.ai API Platform
• GLM-5.2 Documentation
• GLM 5.2 on OpenRouter

Setup

npm install @llmrtc/llmrtc-provider-zai
# or use the backend package, which re-exports it
npm install @llmrtc/llmrtc-backend

Get an API key from the Z.ai platform and set it:

ZAI_API_KEY=your-key

Usage

import { LLMRTCServer, ZaiLLMProvider, OpenAIWhisperProvider, OpenAITTSProvider } from '@llmrtc/llmrtc-backend';

const server = new LLMRTCServer({
  providers: {
    llm: new ZaiLLMProvider({
      apiKey: process.env.ZAI_API_KEY!,
      model: 'glm-5.2' // default
    }),
    stt: new OpenAIWhisperProvider({ apiKey: process.env.OPENAI_API_KEY! }),
    tts: new OpenAITTSProvider({ apiKey: process.env.OPENAI_API_KEY! })
  },
  systemPrompt: 'You are a helpful voice assistant.'
});

await server.start();

CLI mode

LLM_PROVIDER=zai        # or 'glm'
ZAI_API_KEY=your-key
ZAI_MODEL=glm-5.2       # optional override

Configuration

Option	Type	Default	Description
apiKey	string	required	Z.ai API key
model	string	'glm-5.2'	GLM model name
baseURL	string	'https://api.z.ai/api/paas/v4'	API endpoint (Z.ai coding-plan subscribers can point at the coding endpoint)

Tool calling

GLM 5.2 supports the full LLMRTC tool-calling flow — provider-agnostic ToolDefinitions, parallel calls, streamed tool-call assembly, and playbook two-phase execution — through the same OpenAI-compatible adapter used by OpenRouter and LM Studio:

import { ToolRegistry, defineTool } from '@llmrtc/llmrtc-backend';

const registry = new ToolRegistry();
registry.register(defineTool(
  {
    name: 'get_weather',
    description: 'Get the current weather for a city',
    parameters: {
      type: 'object',
      properties: { city: { type: 'string' } },
      required: ['city']
    }
  },
  async ({ city }) => ({ city, temperature: 22 })
));

const result = await llm.complete({
  messages: [{ role: 'user', content: 'Weather in Tokyo?' }],
  tools: registry.getDefinitions(),
  toolChoice: 'auto'
});

Via OpenRouter

If you already use OpenRouter as a gateway, GLM models are available there without a separate Z.ai key:

import { OpenRouterLLMProvider } from '@llmrtc/llmrtc-backend';

const llm = new OpenRouterLLMProvider({
  apiKey: process.env.OPENROUTER_API_KEY!,
  model: 'z-ai/glm-5.2'
});