DSH-Plugin | Models Reasoning for DeepSeek Harness Plugin
Free Models & Reasoning DSH-Plugin for DeepSeek Harness: browse, search, download, install third-party models, routing, cost. Verified, traceable sources.
DSH-Plugin Hub is a community plugin marketplace for DeepSeek Harness (DSH plugin), built to the official plugin spec and embedded in Settings → Plugin Center. Browse, search and one-click install 4000+ human-curated community plugins without leaving the app, synced daily with dsh-plugin.org.
Use ChatGPT, Claude, and Grok subscriptions as DeepSeek Harness LLM providers with OAuth login in the web UI, no API keys.
DeepSeek Harness（DSH）一只住在 DSH 界面右下角的小鲸鱼娘，帮你盯着DeepSeek账户余额。QQ弹弹，支持拖拽吸附、左吸附翻转、数字滚动动画，随界面自动启用，建议直接喊来你的dsh安装
DSH plugin for session cost tracking and OpenCode Go quota display, with budget, heatmap, and official price sync.
DSH plugin attributes token usage to relay sites, zero-config balance and cost estimation.
A macOS-native account-pool and provider plugin for DeepSeek Harness, unifying OAuth sessions and model access.
DSH plugin integrating Command Code as an LLM provider with live model catalog, reasoning-effort support, and Models-page card.
DSH plugin with game-style damage pulse animations for real-time DeepSeek token balance and usage monitoring.
A wallet chip for DeepSeek Harness: live balance, per-session spend & token tracking, low-balance alerts, and one-click recharge.
A DSH plugin for DeepSeek balance monitoring, usage stats, top-up link, low-balance alerts, and third-party plugin manager.
Use ChatGPT subscription to access Codex models in DeepSeek Harness without OpenAI API key.
All-in-one DSH plugin for provider login, model switching, vision fallback, and token/cost analytics.
Connect ChatGPT OAuth and Codex models to DeepSeek Harness with native approvals and session recovery.
DSH plugin tracks token usage and costs per call, with peak/off-peak billing, balance query, and CSV/JSON/PNG export.
Adds a model advanced configuration page to DeepSeek Harness Web UI for custom endpoints and capability settings.
Persistent dock and fully customizable balance/usage panel for DeepSeek Harness, with activity heatmap and dual-channel comparison, local-only.
Native Rapid-MLX provider for DeepSeek Harness, auto-reading model facts from server to eliminate manual config.
Adds Google Antigravity model access to DeepSeek Harness with multi-account pool, 429 rotation, and quota dashboard.
A DSH plugin showing DeepSeek account balance and live session cost in the Web GUI.
Use ChatGPT/Codex subscription directly in DeepSeek Harness without API key, with OAuth login and quota display.
How to pick a trustworthy Models & Reasoning DSH plugin
A Models & Reasoning DSH plugin plugs into and tunes the DeepSeek Harness thinking engine — third-party model adapters, routing strategy, inference parameters and cost control, the brain layer of the plugin ecosystem. It decides which brain your agent thinks with, so check model compatibility, routing and cost control, and compatibility status to balance capability against budget.
- Model compatibility: confirm the plugin supports your model providers and inference protocols (e.g. OpenAI-compatible APIs) and that your DSH version matches.
- Routing and cost: whether it routes tasks to different models, enforces usage limits and falls back on failure — this drives cost and stability.
- Compatibility status: model plugins couple to DSH's inference interfaces — prefer verified plugins whose dshTarget matches to avoid reasoning-chain issues.
Where Models & Reasoning sits in the DSH taxonomy
Models & Reasoning is one of the eleven primary DeepSeek Harness categories, owning the thinking engine: third-party model adapters, routing strategy, inference parameters and cost control. It is the brain layer of the whole plugin ecosystem — Sessions & Messages, Skills & Agents and Tools & Capabilities all run on top of it, so model choice sets the ceiling for every downstream capability.
「Every part of the product is a plugin — including the model adapter, the tool registry, the session log, and the agent loop itself.」
SourcesOfficial site·Architecture docs·Repository