Plugins

DeepSee: DeepSeek Harness vision, multi-model routing, and Gemini visual self-checks before delivery

Plug-in vision for text-only DeepSeek Harness (dsh) models: a `vision` tool with built-in cheap/free VLM presets, multi-image batch analysis, paste-to-hint image admission, and a web settings page with hot-reload.

Scene-awareness probe for DSH on macOS: a privacy-first, pull-model `look` tool that reads the frontmost non-self window (app info, AX title/selection, gated local OCR), plus a summon-hotkey snapshot captured the instant you press.

Personal-experience memory for DSH: turns solved tasks into sourced problem cases with condition-scoped claims, recalls them into new tasks via hidden pre-step injection, and makes corrections/deletes cascade to future recall.

Replaces the chat input model picker with a provider-grouped, collapsible list: groups models by provider, folds vision-bridge mirror providers back into their upstream, and carries the reasoning-effort options each model declares in settings. Tested on DSH Desktop only; the DSH web build is untested.

Memory archive and prompt viewer for DeepSeek Harness: read back what compaction folded out of context - browse past sessions or Tavern playthrough archives (summary, original text, state) with full-text search - and inspect the prompt each turn actually sent, listing every assembled system section with its order, character count, offset, hash and mutability, plus an assembly map, registry export, and honest diagnostics when a section is empty or its log cannot be read. Optionally provisions a roleplay agent preset.

Large-image recognition for vision-exp models: lossless 800×800 tile recognition (smart/pipeline/full), local OCR with preprocessing & handwriting routing, optional multi-vendor GPU (DirectML/CUDA/OpenVINO) with auto CPU fallback.

DeepSeek Harness skill: build, rebuild and refresh a course module's revision atlas — a mind map whose leaves are paged notebooks, extracted faithfully from the module's markdown with machine-checked coverage and a human-gated plan.

Automatic tier-based model routing for DeepSeek Harness: a virtual `smart` model classifies each request by difficulty (hard / normal / easy) and by vision need, then delegates it to the models you already configured. 三级难度 + 视觉自动路由，虚拟 smart 模型零配置接入。

Turn the DeepSeek Harness Web UI into a real desktop application from inside DSH: one plugin that provisions a portable launcher (native WebView2 window, tray residency, silent backend start, readiness gating, debug console) plus desktop/start-menu shortc

Xiaomi MiMo-powered voice for DeepSeek Harness: browser 🎤/🧠/🔊 UI, voice_transcribe/voice_understand/voice_speak tools, configurable voice map (preset/voicedesign/voiceclone). Fork of zhuiyueya/dsh-voice (MIT), Settings pattern from Anionex/dsh-vision-toolkit (MIT).

Image and video production toolkit for DeepSeek Harness - background removal with quality tiers, local inpainting retouch, headline and logo compositing, batch platform-size exports and image QA reports; for video, AI subject matting with background replacement (RVM ONNX, plus a chroma-key path), animated titles with fade/slide/typewriter/karaoke presets, 53 transitions with stylized variants, and light-effect and grading presets with LUT support, alongside footage inventory, accurate cutting, subtitle burn-in, music mixing and platform-spec exports; it provisions its own local Python runtime through a bundled bootstrap script.

Digests the images in a prompt with a vision model before admission, so any model — a text-only one included — can read them without the session ever switching models, and adds a describe_image tool for image paths.

Trims historical images in outgoing chat requests down to a recent-image count, learns the provider per-prompt image cap from HTTP 400 responses, and retries with fewer images so image-heavy sessions keep working.

Peer communication between conversations in one DeepSeek Harness process: list the conversations the workspace shows, link one under a nickname, message it without waiting for its turn to end, read its progress without waking it, and declare rules for your own tool calls.

Auto-switch DeepSeek models: routes every request to the vision model when images are present, to the pro model for complex tasks, and to the fast model otherwise; includes a switch_model tool for manual overrides.

Token-cost pipeline for DeepSeek Harness: pressure-aware layered compaction (lossless projection below 40% context pressure, LLM summarization above), per-agent tool allowlisting, long-text-to-image with vision summarization, and a /token-status command tracking pressure, tiers, and saved tokens.

Whole-book translation production line for DeepSeek Harness: deterministic chapter detection and segmentation, per-chapter subagent translation, glossary control, full-book review, targeted revision, and auditable meta.json evidence.

Evidence-first attention supervision for DeepSeek Harness with C0–C3 policy, deduplicated Inbox alerts, quiet hours, and redacted outcome records.

Vision bridge for text-only DeepSeek models: transcribes attached images with deepseek-v4-flash-vision-exp before they reach deepseek-v4-pro, with no third-party dependencies.

Apple on-device Vision tools for text-only dsh models: local OCR (zh-Hans + 30 langs), image classification, face detection, document layout, and a combined describe — 100% offline, no API key, tall-screenshot slicing.

Adds image input and recognition through configured DSH providers or an OpenAI-compatible endpoint.

Windows desktop automation for DeepSeek Harness: accessibility-first observe/act loop with 30 standard tools (element/coordinate targets, auto/a11y/event strategy routing), superseded-observations and no-replay safety semantics, persistent kill switch, RuntimeId/rect-drift stale protection, chat cards and a floating panel (live ops, frame wall, env auto-detect settings), tiled vision describe for low-res models, JSONL audit trail and a pid allowlist. Requires Windows 10+ and Python 3.9+ with uiautomation and pillow (installable from the settings page).

Auto-switch the DeepSeek route to the vision model on demand: flash main session switches (A), pro keeps deep reasoning and delegates image reading to a vision subagent (B), subagents always switch, with fatal-failure fallback. No manual model switching.