z.ai provider

z.ai is API-token based. No browser cookies.

Token sources (fallback order)

1. Config token (~/.codexbar/config.json → providers[].apiKey).
2. Environment variable Z_AI_API_KEY.

Config location

• ~/.codexbar/config.json

API endpoint

• GET https://api.z.ai/api/monitor/usage/quota/limit
• BigModel (China mainland) host: https://open.bigmodel.cn
• Override host via Providers → z.ai → API region or Z_AI_API_HOST=open.bigmodel.cn.
• Override the full quota URL (e.g. coding plan endpoint) via Z_AI_QUOTA_URL=https://open.bigmodel.cn/api/coding/paas/v4.
• Headers:
  • authorization: Bearer <token>
  • accept: application/json

Parsing + mapping

• Response fields:
  • data.limits[] → each limit entry.
  • data.planName (or plan, plan_type, packageName) → plan label.
• Limit types:
  • TOKENS_LIMIT → primary (tokens window).
  • TIME_LIMIT → secondary (MCP/time window) if tokens also present.
• Window duration:
  • Unit + number → minutes/hours/days.
• Reset:
  • nextResetTime (epoch ms) → date.
• Usage details:
  • usageDetails[] per model (MCP usage list).

Key files

• Sources/CodexBarCore/Providers/Zai/ZaiUsageStats.swift
• Sources/CodexBarCore/Providers/Zai/ZaiSettingsReader.swift
• Sources/CodexBar/ZaiTokenStore.swift (legacy migration helper)