AI Agent Hub

DeepSeek Harness Plugins · Model Inference

Plugins that extend DeepSeek Harness (DSH) with vision, workflows, memory, web tools and more.

🤖
dsh-zpdf
komoai2026/dsh-zpdf

ZPDF tools for DeepSeek Harness: high-fidelity PDF-to-Markdown parsing, layout-preserving translation, multi-format export, and credit management.

Model Inference 2026.08.23 ★ 1
dsh-visibridge preview
dsh-visibridge
lhbsaa/dsh-visibridge

A host-level plugin for DeepSeek Harness that adds image recognition and live camera capture to pure-text models, supporting local Ollama and any OpenAI-compatible vision backend with structured JSON evidence.

Model Inference 2026.08.25 ★ 1
🤖
dsh-subscriptions
DevViking-Persike/dsh-subscriptions

Use your own Claude and ChatGPT/Codex subscriptions as model providers in DeepSeek Harness via OAuth sign-in, no API key required, with text and image input support.

Model Inference 2026.08.25 ★ 1
🤖
aitoubiaoling-bid-review
zslzxy/aitoubiaoling-bid-review

AI-powered bid reviewer that checks commercial, technical, and attachment compliance against tender requirements, producing a traceable Markdown risk report.

Model Inference 2026.08.25 ★ 1
dsh-pianist preview
dsh-pianist
Laplace-bit/dsh-pianist

An AI piano player for DeepSeek Harness: request a song in natural language, let the model compose and perform it with realistic sampling and an immersive concert stage.

Model Inference 2026.08.25 ★ 1
🤖
dsh-auto-chat-titles
leonardoxr/dsh-auto-chat-titles

Generates semantic LLM-based chat titles for DeepSeek Harness with model selection, one-click regeneration, and configurable length and context settings.

Model Inference 2026.08.25 ★ 1
dsh-live-token-stats preview
dsh-live-token-stats
better-er/dsh-live-token-stats

Real-time token rate and first-token latency monitoring for DeepSeek streaming responses, with built-in BPE-based token estimation and stall detection.

Model Inference 2026.08.27 ★ 1
🤖
vibe-intent-compiler
Shhaaawwww/vibe-intent-compiler

One-click rewrite of messy drafts in the DeepSeek Harness composer into concise, faithful, actionable instructions without inventing details.

Model Inference 2026.08.25 ★ 1
🤖
dsh-qwen38-compaction-fix
zhubaohi/dsh-qwen38-compaction-fix

Fixes truncated compaction checkpoints for local qwen3.8-27b on NInfer by disabling thinking and tuning sampling parameters for compaction calls only.

Model Inference 2026.08.25 ★ 1
dsh-prompt-enchant preview
dsh-prompt-enchant
Kian-Oraish/dsh-prompt-enchant

Adds a magic-wand button to the DSH web input that enhances casual prompts into precise expressions via a dedicated LLM call, with General and Design modes, reference protection, and confirmation dialogs.

Model Inference 2026.08.26 ★ 1
dsh-image-gen preview
dsh-image-gen
Anyway-one/dsh-image-gen

Adds text-to-image and image-to-image generation to DeepSeek Harness via any OpenAI Images-compatible endpoint, requiring only a baseURL and API Key.

Model Inference 2026.08.25 ★ 1
🤖
dsh-camel
JuwanXu/dsh-camel

Opt-in rate-limit recovery for free models with adaptive quota learning, proactive request pacing, and configurable retries that keep tasks alive instead of terminating on throttling.

Model Inference 2026.08.26 ★ 1