| summary | read_when | ||
|---|---|---|---|
z.ai provider data sources: API token in config/env and quota API response parsing. |
|
z.ai is API-token based. No browser cookies.
- Config token (
~/.codexbar/config.json→providers[].apiKey). - Environment variable
Z_AI_API_KEY.
~/.codexbar/config.json
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
- Response fields:
data.limits[]→ each limit entry.data.planName(orplan,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).
Sources/CodexBarCore/Providers/Zai/ZaiUsageStats.swiftSources/CodexBarCore/Providers/Zai/ZaiSettingsReader.swiftSources/CodexBar/ZaiTokenStore.swift(legacy migration helper)