dsh-llm-pi-ai-live

Real-time model-list refresh for DSH llm-pi-ai routes: queries each provider endpoint, merges the installed pi-ai catalog metadata onto the live ids, and appends new models through the official settings seam. 适合需要动态同步和更新各服务商最新模型列表的系统管理员。

パッケージ

dsh-llm-pi-ai-live

互換性

未検証

Harness ピア範囲

>=0.1.7-rc.2 <0.2.0

Cordis ピア範囲

>=4.0.4 <5

バージョン

0.1.0

ライセンス

MIT

最終更新

2026/09/29

Configuration

Every field is optional; defaults are documented in cordis.patch.yml.

Field	Default	Meaning
`enabled`	`true`	Master switch; `false` loads the plugin but schedules nothing
`startupDelayMs`	`10000`	Delay after mount before the first pass (ms); `0` runs it immediately, and the startup pass always runs
`intervalMs`	`21600000`	Period between passes (ms, six hours); `0` disables the periodic pass
`timeoutMs`	`30000`	Per-request network timeout
`settingsNamespaces`	`[]`	Namespaces to serve; empty auto-detects the pi-ai namespace
`include`	`[]`	Routes to refresh; empty means every configured route
`exclude`	`[]`	Routes to leave alone, e.g. `['openrouter']`
`enrichFromCatalog`	`true`	Fill new entries from the installed pi-ai catalog
`maxModels`	`2000`	Cap on stored model entries per route
`dryRun`	`false`	Compute and report without writing
`toolEnabled`	`true`	Register the `refresh_model_catalog` tool

dsh-llm-pi-ai-live

インストール

ドキュメント

Configuration