--- name: zai-setup description: > Use this skill when the user asks about z.ai, Zhipu AI, setting up z.ai API access, getting a z.ai API key, configuring Claude Code to use z.ai models directly via the Anthropic-compatible endpoint, routing z.ai through Cloudflare AI Gateway or OpenRouter, or using z.ai's OpenAI-compatible API. --- # z.ai (Zhipu AI) Setup Guide [z.ai](https://z.ai/) (formerly Zhipu AI / 智谱AI, rebranded July 2025) is a Chinese AI company that develops the GLM (General Language Model) family. They provide both OpenAI-compatible and Anthropic-compatible API endpoints. - **Platform**: (international) / (China) - **Docs**: - **API Base URL (International)**: `https://api.z.ai/api/paas/v4/` - **API Base URL (China)**: `https://open.bigmodel.cn/api/paas/v4/` - **Anthropic-compatible endpoint**: `https://api.z.ai/api/anthropic` - **Pricing**: ## Getting an API Key 1. Visit (international) or (China) 2. Create an account (email or phone verification required) 3. Navigate to **API Keys** in your account dashboard 4. Generate and copy your API key ## API Compatibility z.ai provides **two** compatible API formats: ### OpenAI-Compatible API ```bash curl "https://api.z.ai/api/paas/v4/chat/completions" \ -H "Authorization: Bearer YOUR_ZHIPU_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "glm-4.7", "messages": [{"role": "user", "content": "Hello!"}] }' ``` ```python from openai import OpenAI client = OpenAI( api_key="YOUR_ZHIPU_API_KEY", base_url="https://api.z.ai/api/paas/v4/" ) response = client.chat.completions.create( model="glm-4.7", messages=[{"role": "user", "content": "Hello!"}] ) ``` ```typescript import OpenAI from "openai"; const client = new OpenAI({ apiKey: "YOUR_ZHIPU_API_KEY", baseURL: "https://api.z.ai/api/paas/v4/", }); const response = await client.chat.completions.create({ model: "glm-4.7", messages: [{ role: "user", content: "Hello!" }], }); ``` ### Anthropic-Compatible API z.ai also provides an endpoint that speaks the Anthropic Messages API protocol. This is what enables direct Claude Code integration. ## Using z.ai with Claude Code ### Option 1: Direct via Anthropic-Compatible Endpoint (Recommended) z.ai provides a native Anthropic-compatible endpoint at `https://api.z.ai/api/anthropic`. Claude Code can connect directly — no proxy needed. **In `~/.claude/settings.json`:** ```json { "env": { "ANTHROPIC_BASE_URL": "https://api.z.ai/api/anthropic", "ANTHROPIC_AUTH_TOKEN": "your-zai-api-key", "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1" } } ``` **Or via shell alias:** ```bash # Add to ~/.bashrc or ~/.zshrc zaiclaude() { ANTHROPIC_BASE_URL=https://api.z.ai/api/anthropic \ ANTHROPIC_AUTH_TOKEN="your-zai-api-key" \ CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1 \ claude "$@" } ``` **Optional model mapping** — map Claude Code's model slots to specific GLM models: ```json { "env": { "ANTHROPIC_BASE_URL": "https://api.z.ai/api/anthropic", "ANTHROPIC_AUTH_TOKEN": "your-zai-api-key", "ANTHROPIC_DEFAULT_HAIKU_MODEL": "glm-4.5-air", "ANTHROPIC_DEFAULT_SONNET_MODEL": "glm-4.7", "ANTHROPIC_DEFAULT_OPUS_MODEL": "glm-5", "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1" } } ``` | Variable | Purpose | | ------------------------------------------ | ---------------------------------------------------------- | | `ANTHROPIC_BASE_URL` | Points Claude Code to z.ai's Anthropic endpoint | | `ANTHROPIC_AUTH_TOKEN` | Your z.ai API key | | `ANTHROPIC_API_KEY` | Set to `""` to avoid conflicts with existing Anthropic key | | `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` | Prevents background traffic to Anthropic servers | | `ANTHROPIC_DEFAULT_SONNET_MODEL` | GLM model for the "Sonnet" slot | | `ANTHROPIC_DEFAULT_OPUS_MODEL` | GLM model for the "Opus" slot | | `ANTHROPIC_DEFAULT_HAIKU_MODEL` | GLM model for the "Haiku" slot | > **Note**: See for z.ai's official Claude Code integration guide. ### Option 2: Via Cloudflare AI Gateway Route z.ai through [Cloudflare AI Gateway](https://developers.cloudflare.com/ai-gateway/) for logging, caching, and cost tracking. See the `cloudflare-ai-gateway` skill for full setup. ```bash # Use the universal endpoint to proxy z.ai through your gateway curl "https://gateway.ai.cloudflare.com/v1/ACCT_ID/GATEWAY_ID" \ -H "Content-Type: application/json" \ -d '[{ "provider": "zhipu", "endpoint": "https://api.z.ai/api/paas/v4/chat/completions", "headers": { "Authorization": "Bearer YOUR_ZHIPU_API_KEY" }, "query": { "model": "glm-4.7", "messages": [{"role": "user", "content": "Hello"}] } }]' ``` ### Option 3: Via OpenRouter [OpenRouter](https://openrouter.ai/) aggregates many model providers including z.ai/Zhipu models: ```bash curl "https://openrouter.ai/api/v1/chat/completions" \ -H "Authorization: Bearer $OPENROUTER_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "zhipu/glm-4.7", "messages": [{"role": "user", "content": "Hello"}] }' ``` ### Claude Code Web Sessions For Claude Code web sessions, set environment variables in `.claude/settings.local.json` (gitignored): ```json { "env": { "ANTHROPIC_BASE_URL": "https://api.z.ai/api/anthropic", "ANTHROPIC_AUTH_TOKEN": "your-zai-api-key", "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1" } } ``` > **Important**: Never put API keys in `settings.json` (committed). Always use `settings.local.json` (gitignored) or a secrets manager like 1Password. These use the same Anthropic-compatible environment variables as Option 1 above. ## Coding Plan z.ai offers a **Coding Plan** subscription for enhanced coding model access: - **Lite**: $10/mo - **Pro**: $30/mo - **Endpoint**: `https://api.z.ai/api/coding/paas/v4` ## References - [z.ai Platform](https://z.ai/) - [z.ai Developer Docs](https://docs.z.ai/) - [z.ai Pricing](https://docs.z.ai/guides/overview/pricing) - [Claude Code Integration Guide](https://docs.z.ai/scenario-example/develop-tools/claude) - [OpenRouter Zhipu Models](https://openrouter.ai/models?q=zhipu) - [Cloudflare AI Gateway Universal Endpoint](https://developers.cloudflare.com/ai-gateway/providers/universal/) - [z.ai on GitHub](https://github.com/zai-org) - [z.ai on Hugging Face](https://huggingface.co/zai-org)