dsh-anyrouter
Run the following command in DeepSeek Harness:
dsh plugin install shaomingbo/dsh-anyrouter
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install shaomingbo/dsh-anyrouter inside DeepSeek Harness to install the plugin; the source lives at https://github.com/shaomingbo/dsh-anyrouter . Restart DSH manually and hard-refresh the Web page afterwards to activate it.
About this plugin
Routng Claude and GPT/Codex through the Any Router relay demands exact request identity — transport format, beta headers, Agent SDK identity blocks — or the relay simply rejects the call. dsh-anyrouter is a dedicated DeepSeek Harness provider plugin that sends Claude over a Claude Code 2.1.239-compatible Anthropic Messages transport and GPT/Codex over POST /v1/responses with the Codex CLI request shape, so both families are correctly recognised and relayed upstream.
Synchronising a model persists its full reasoning profile: selectable effort tiers, default effort, the adaptive-thinking flag, and context/output capacities. Both the model selector and the reasoning selector therefore work for every synchronised model, and every value stays user-editable in the picker. Without an API key the entire route goes dormant and the model selector hides the group; once the key is stored the route springs back without a restart. The model list appears as a checkbox picker, so you can adopt a subset and fine-tune each row's reasoning configuration before saving.
If you are running Claude and GPT/Codex behind the Any Router relay and want a single DSH Web panel to manage models and reasoning parameters, this plugin is the purpose-built solution. It is verified end-to-end against DSH 0.1.2-alpha.3, and the API key always lives in the credentials service — the browser can write or replace it but can never read it back, shrinking the key-exposure surface.
Use Cases
- Calling both Claude and GPT/Codex behind an Any Router relay
- Managing the model list and reasoning tiers from the DSH Web panel
- Checking out a model subset and fine-tuning effort per row
Best For
- Users who have deployed an Any Router relay and need DSH integration
- Developers who prefer a Web panel over raw curl for multi-model inference
- Security-conscious setups where the key must be write-only and never read back
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.