harald/zeroclaw
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 3
zeroclaw/docs/zai-glm-setup.md
adisusilayasa 402d8f0a32 docs: add Z.AI GLM coding plan setup guide 
- Add comprehensive documentation for Z.AI GLM models
- Include curl examples for testing Z.AI API
- Document available models and troubleshooting
- Update .env.example with Z.AI configuration

Z.AI provides GLM models (glm-4.5, glm-4.6, glm-4.7, glm-5) through
the OpenAI-compatible endpoint at api.z.ai/api/coding/paas/v4.

Existing tests verify:
- zai_base_url() returns correct URLs for global/CN variants
- create_provider('zai', key) successfully creates provider
- Regional alias predicates cover all variants
2026-02-18 15:10:55 +08:00

2.6 KiB
Raw Blame History

Z.AI GLM Coding Plan Setup

ZeroClaw supports Z.AI's GLM models through multiple endpoints. This guide covers the recommended configuration.

Overview

Z.AI provides GLM models through two API styles:

Endpoint API Style Provider String
/api/coding/paas/v4 OpenAI-compatible zai
/api/anthropic Anthropic-compatible anthropic-custom:https://api.z.ai/api/anthropic

Setup

Quick Start

zeroclaw onboard \
  --provider "zai" \
  --api-key "YOUR_ZAI_API_KEY"

Manual Configuration

Edit ~/.zeroclaw/config.toml:

api_key = "YOUR_ZAI_API_KEY"
default_provider = "zai"
default_model = "glm-4.7"
default_temperature = 0.7

Available Models

Model Description
glm-4.5 Stable release
glm-4.5-air Lightweight version
glm-4.6 Improved reasoning
glm-4.7 Current recommended
glm-5 Latest

Verify Setup

Test with curl

# Test OpenAI-compatible endpoint
curl -X POST "https://api.z.ai/api/coding/paas/v4/chat/completions" \
  -H "Authorization: Bearer YOUR_ZAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-4.7",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Expected response:

{
  "choices": [{
    "message": {
      "content": "Hello! How can I help you today?",
      "role": "assistant"
    }
  }]
}

Test with ZeroClaw CLI

# Test agent directly
echo "Hello" | zeroclaw agent

# Check status
zeroclaw status

Environment Variables

Add to your .env file:

# Z.AI API Key
ZAI_API_KEY=your-id.secret

# Or use the provider-specific variable
# The key format is: id.secret (e.g., abc123.xyz789)

Troubleshooting

Rate Limiting

Symptom: rate_limited errors

Solution:

  • Wait and retry
  • Check your Z.AI plan limits
  • Try glm-4.7 instead of glm-5 (more stable availability)

Authentication Errors

Symptom: 401 or 403 errors

Solution:

  • Verify your API key format is id.secret
  • Check the key hasn't expired
  • Ensure no extra whitespace in the key

Model Not Found

Symptom: Model not available error

Solution:

  • List available models:
curl -s "https://api.z.ai/api/coding/paas/v4/models" \
  -H "Authorization: Bearer YOUR_ZAI_API_KEY" | jq '.data[].id'

Getting an API Key

  1. Go to Z.AI
  2. Sign up for a Coding Plan
  3. Generate an API key from the dashboard
  4. Key format: id.secret (e.g., abc123.xyz789)

Related Documentation