tellmewhattodo/dsh-any-deepseek
One DeepSeek V4.1 Flash provider for any number of third-party relays: paste a key per channel in settings, pick which channel is live, and see token usage booked per real upstream.
Collapses a list of relay endpoints into one provider row, with a channel per relay and one active at a time. Token usage — input, cache-read, cache-write and output — is recorded per request and grouped by the channel that actually served it, so the cost of each relay is visible instead of pooled.
Install
dsh plugin --profile web add dsh-any-deepseekThe npm line from the README; dsh-any-deepseek is published and its registry metadata points back at this repository.
Compatibility
The README states the model list is deliberately exactly one model, so the model picker does not grow a provider group per relay. Only a key is user input: endpoint, upstream model id, context window, output cap and compatibility flags ship in the plugin's own preset table. Keys never enter configuration — they live in the harness credential store and are not written into config files. Channels are distinct from protocols, and the README documents switching channels, thinking level, display names, image input and usage accounting.
Details
- Repo: tellmewhattodo/dsh-any-deepseek
- Category: Models & Providers
- Stars: 0
- Version: 0.0.15 (npm latest)
Alternatives
Jensen-Yao/dsh-model-palette · lcestou/dsh-oh-my-claude · Vmarcelo49/dsh-local-models