Skip to main content

How to use dsh-claude-provider

Add an independent Claude provider type for DSH model settings, with gear levels based on model configuration, and support for Anthropic native model detection with pre-filled known model capacities.

This article is auto-derived from indexed fields (wiki / faq / compatibility_json), not freshly AI-generated.

This article is derived from the plugin's already-indexed fields (wiki / faq / compatibility_json / readme), not freshly generated by AI. Source field is noted at the end of each section.

Quick start

dsh-claude-provider

— source: plugin_wiki.wiki_content

Install & verify

dsh plugin --profile web add @mofeng2223/dsh-claude-provider

Run the command above in your DSH Web Profile. Then enable the plugin in the plugin list.

— source: plugins.install

Key points

    • Five levels: Low, Medium, High, XHigh, and Max
    • Four levels: Low, Medium, High, and Max
    • Toggle: On or Off

— source: plugin_wiki.readme_en (fallback readme_raw)

FAQ

Do I need to manually edit settings.yaml after installation?

No. When the frontend saves a Claude provider, the plugin automatically writes the corresponding native configuration to settings.yaml.

Will it modify DSH's built-in Anthropic provider or regular custom providers?

No. Model probing, thinking effort levels, and capacity pre-filling only apply to Provider IDs explicitly registered as Claude providers; built-in anthropic routes and other anthropic-messages custom routes maintain their original behavior.

Does it support custom Anthropic-compatible gateways?

Yes. Model probing derives the /v1/models endpoint from the Messages API's baseURL, allowing any HTTP(S) address; API keys use saved credential references or are passed directly during calls.

Will uninstalling the plugin lose existing Claude providers?

No. Uninstalling does not delete settings.yaml or credentials; saved providers continue as regular anthropic-messages routes, and native fields like reasoningEfforts, reasoning, and compat.forceAdaptiveThinking continue to work, only losing the dedicated UI, model probing, and capacity pre-filling.

What thinking effort levels are available for each model?

Three presets are available: five levels (low/mid/high/ultra-high/max) suitable for Opus 5, Sonnet 5, etc.; four levels (low/mid/high/max) suitable for Opus 4.6, Sonnet 4.6; toggle (on/off) suitable for Haiku 4.5. New models default to five levels, and Claude providers default to high.

Can it be used in command-line mode (without web interface)?

Yes, change --profile to headless in the installation command, but model probing and the dedicated UI require a web client to run.

— source: plugin_wiki.faq_json

Compatibility

  • DSH: 未声明
  • Node: >=22.19

— source: plugin_wiki.compatibility_json

Pitfalls

Review the upstream repo before installing. This guide is auto-derived from indexed fields and may lag the latest release. If anything contradicts the official docs, treat the upstream source as authoritative.

— source: general rule