dsh-llm-kimi
Run the following command in DeepSeek Harness:
dsh plugin install haveanote06/dsh-llm-kimi
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install haveanote06/dsh-llm-kimi in your terminal to install; find the source at https://github.com/haveanote06/dsh-llm-kimi.
About this plugin
The DeepSeek Harness ecosystem previously lacked a native adapter for Kimi and Moonshot models. Routing Kimi inference through Harness often meant hand-rolling an OpenAI-compatible bridge or bypassing the router entirely. dsh-llm-kimi plugs this gap as a first-class LLM adapter, letting Harness speak Kimi request payloads, SSE streams, and function-calling semantics directly.
Three routes are provided: Kimi Code subscription, and the Moonshot open platform for both CN and global endpoints. The model catalog spans from moonshot-v1-8k up to k3 with a 1M context window. Thinking mode offers three reasoning-effort tiers (low, high, max), function calling and base64 multimodal image input are verified against the live API, and credential or route changes take effect on the very next request with no restart required. A dedicated Kimi settings page ships with the plugin for day-to-day configuration.
Best suited for developers who already build workflows on DeepSeek Harness and want to fold Kimi into a unified scheduling layer, as well as research and engineering teams that need tool-calling orchestration, multimodal understanding, or high-reasoning tasks across multiple model families.
Use Cases
- Orchestrate Kimi and DeepSeek models within a single Harness workflow
- Leverage Kimi thinking mode for high-effort code generation and analysis
- Process multimodal image inputs for visual understanding and code review
Best For
- Developers building workflows on DeepSeek Harness
- Engineering teams orchestrating function calls across multiple models
- Researchers integrating Kimi into unified inference pipelines
Related Plugins
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.
An armor-breaking plugin for DeepSeek: rewrites instruction priority, expands refusal-blocked phrases, persists across sessions, and shows a green active indicator.