跳到主要内容

dsh-provider-timeweb

已验证

dsh-provider-timeweb · v0.4.0 · MIT · Web 界面

TimeWeb Cloud AI provider for DeepSeek Harness — a "Time Web" section with "Добавить агента Time Web": auto-fetch models from the agent, one-click create of a correct provider, success toast, auto-close to Models.

安装

dsh plugin add dsh-provider-timeweb

dsh --profile default --dump-config 确认 layer 已生效 —— 参见安装指南

源码

发布到 npm 但没有公开仓库。安装前请检查包内容。

标签

说明文档

dsh-provider-timeweb — TimeWeb Cloud AI provider for DeepSeek Harness

Adds a Time Web section to DSH Settings with a ready-made «Добавить агента Time Web» form:

  • Автоподхват моделей — press «Подтянуть модели» and DSH fetches the agent's model list from TimeWeb automatically.
  • Один клик — the form writes a correct, working provider (api: openai-completions, required x-proxy-source header, right limits).
  • Подтверждение — after saving, a «✓ Всё хорошо» toast shows, then the panel returns to the Models section.

No manual edits to settings.yaml, no fiddling with the provider catalog.

Install

dsh plugin --profile web add dsh-provider-timeweb
dsh web

Add a TimeWeb agent (all through the UI)

  1. Open Settings → Time Web.
  2. Press + Добавить агента Time Web.
  3. Fill in:
    • Отображаемое имя — e.g. TimeWeb
    • Идентификатор (route) — lowercase, e.g. timeweb2
    • Access ID агента — from the agent card in TimeWeb
    • Токен агента — JWT from «Токены авторизации»
    • proxy-source — leave dsh-harness
  4. Press Подтянуть модели to fetch the agent's model list, or keep the default yandex/yandexgpt-pro-5.1 (context 32000, max tokens 4096).
  5. Press Добавить.

The provider lands in settings.yaml (llm-pi-ai.providers.<route>); the token is stored via the credentials service (never in settings.yaml). Pick the model via /model and chat.

Repeat for more agents — each with its own route, Access ID, and token.

Removing

dsh plugin --profile web remove dsh-provider-timeweb

Notes

  • Requires DSH 0.1.1-rc.2 and Node.js 22+.
  • The plugin only adds a settings section; the standard Settings → Models flow stays exactly as shipped.