DeepSeek Harness Plugins
DeepSeek Harness PluginManifest Valid

dsh-local-llm-controller

Control and launch local llama.cpp instances directly from DSH settings to serve local LLMs in chat.

VisionDeveloper Tools
4GitHub Stars0ForksUpdated2026-09-17

Install

$ dsh plugin --profile web add dsh-local-llm-controller

Plugin Overview & Capabilities

AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.

source-grounded
dsh-local-llm-controller is a DeepSeek Harness plugin providing a dedicated UI card under Settings → Plugins to manage local llama.cpp `llama-server` instances. It supports dual model slots (A/B), custom launch parameter presets across text/vision and fast/long context modes, automatic GGUF model registration into DSH model providers, and lifecycle control without manual terminal commands.

Key Capabilities

  • One-click start and stop of local llama.cpp llama-server processes from DSH Settings
  • Dual slot architecture (Slot A/B) supporting independent model directories and GGUF files
  • Eight customizable launch parameter sets across text/vision modes and fast/long-context presets
  • Automatic discovery and mounting of mmproj projection files for multimodal vision models
  • Automatic registration of local model providers and names derived from chosen GGUF files into DSH
  • Process lifecycle cleanup on DSH shutdown or restart

Useful For

  • Running local open-source LLMs (like Qwen MoE or Dense models) offline within DeepSeek Harness
  • Switching dynamically between fast text generation and extended long-context or multimodal vision tasks
  • Managing llama-server arguments, VRAM offloading, and context windows directly from the web interface

Who It Fits

  • Developers and power users running local llama.cpp setups with DeepSeek Harness
  • Privacy-conscious users hosting GGUF models locally on private hardware

Documented Limitations

  • Requires an existing local llama.cpp installation (llama-server executable) and pre-downloaded GGUF models
  • Older llama-server builds do not support WebP images, requiring JPEG/PNG conversions for vision input
  • Switching model files on the same slot does not automatically delete old provider entries in DSH model settings

DSH Compatibility

Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.

Not tested yetNo runtime compatibility tests have been published yet.

Security Signals

Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.

No automated security signals have been published yet.

Source & Registry Notes

Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.

Source repository
Lbunc/dsh-local-llm-controller
Registry source
GitHub · dsh-plugin topic
Registry classification
Plugin
Source checked
87f23f1 · 2026-09-18

This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.

Repository Activity

GitHub stars
4GitHub stars
Forks
0
Open issues
0
Last commit
2026-09-17
Last release
2026-09-10
For maintainers

Maintaining this plugin?

This listing is generated from public repository data. If you maintain this project, you can review the information and share this listing with your users if you find it useful.

Add to README
Listed on DSHPlugin.app