Awesome DeepSeek Harness Plugin

Vision bridge for text-only models: paste an image, get structured JSON evidence (OCR, layout, semantics).

Free vision for text-only agents: built-in keyless vision chain plus pixel tools (Q&A, grounding, crop, pixel diff, colors, OCR, SVG trace, cutout, screenshots); paste an image to use it.

Turns a video (a local file, or a Bilibili/yt-dlp link) into one addressable storyboard grid — numbered cells with exact timestamps, per-cell transcript and an audio envelope on the same timeline — and video_zoom re-reads any window at higher resolution, so a text-only model can look at the video instead of reading a summary.

Lets a text-only agent watch a live screen or camera feed: every frame is scored by how much the picture changed, only the ones that matter are packed into one numbered sheet with an audio envelope, so the model looks at what actually changed instead of a screenshot every few seconds.

Vision for text-only models: paste an image and the model switches to a Vision Toolkit variant for image Q&A, multi-image comparison, long-screenshot OCR, screenshot-to-UI reproduction, element grounding, and pixel diff. No API key by default — images are processed by the author-hosted free service, 100 per machine per day; configurable to your own provider.

AI image generation for the DSH Web GUI: text-to-image and image-to-image through a configurable OpenAI-compatible endpoint (gpt-image-2 / gpt-image-1 / dall-e-3), with an api_url/api_key settings card and a sidebar split-pane generation studio.

AI image and video generation for the DSH Web GUI: text-to-image and image-to-image over per-channel model catalogs with per-family request shaping (gpt-image, Grok Imagine, Nano Banana, Seedream, GLM-Image, Qwen-Image, MiniMax image-01, DALL-E 3), a sidebar generation studio, an infinite canvas with node workflows, files preview and canvas skills, and a prompt template library; the video side adds text-to-video, image-to-video, first/last-keyframe and multi-reference modes on an async submit/poll/download pipeline, host-side video tools, video nodes on the canvas, and clip stitching of two or more video nodes into one MP4 with the local ffmpeg.

Drive a local or remote ComfyUI server from DeepSeek Harness: comfyui_run / comfyui_object_info / comfyui_workflow tools generate and edit images and videos, with a workflow library (graph extraction: per component / main flow / all), a load area with resolution auto-match, a live queue, SDXL and Wan 2.1 templates, a companion skill, and a same-origin media proxy.

Upgrades DSH image viewing with pointer-centered zoom, pan, galleries, downloads, keyboard support, and inline region notes.

Lets text-only models handle pasted chat images, with a native vision experience, batch image viewing, and a built-in OpenAI-compatible analyze_image tool; vision-capable models are unaffected.

Reads animated GIF and WebP as animations: a pasted one keeps playing in the UI while the model gets a labelled sheet of frames sampled across it, and the read_animation tool samples a file on demand.

Image generation and image reading for DSH: an Images page in the sidebar, labelled 图像 in Chinese (one row per model, 10 ratios, 1K–4K sizes, a contact sheet with detail cards, zoom, reuse and delete), three preset providers plus custom endpoints — one needs no key at all — and a vision-review skill. A pasted key is written to ~/.dsh/secrets/media-tools.env (mode 600) and only sent to the provider you picked. The page follows the harness language (Chinese or English). Requires Python 3.9+ (generation runs through scripts/generate.py); the page needs desktop DSH 0.2.1+.

Send local images or HTTP(S) image URLs to a configured OpenAI-compatible vision endpoint with the inspect_image tool and return its text response to the conversation.

Multi-engine text-to-image generation (OpenAI Images and Zhipu CogView presets) with per-session quota tracking, engine failover, credential-safe config, and a result card with regenerate.

AI image and video studio as a floating DSH panel, so creation runs alongside the conversation instead of replacing it: text-to-image, image-to-image and multi-image composition at 1K-4K across eight aspect ratios; text-to-video and image-to-video with first-frame control at 4-12 seconds; short-drama mode that imports a script (.txt/.md/.json), breaks it into storyboard shots for preview and batch generation; and a prompt-expert workspace. Zero runtime dependencies.

Local video understanding for agents: extracts keyframes adaptively based on motion and pauses, inspects any timestamp at full resolution, and aligns visuals with dialogue from embedded subtitles or configured ASR. Requires ffmpeg and ffprobe; clips without subtitles also need a transcription engine for dialogue (a ~190 MB local whisper model, a cloud API key, or the official voice-input model).

Model-facing `vision` tool for DeepSeek Harness: describe and OCR image files by calling the free Zhipu GLM vision API directly (glm-4v-flash fallback chain), no external CLI required.

Samples a video the way a person does instead of downloading it: metadata and chapters first, then timestamped frames as images, subtitle-first transcripts (local ASR only when the platform carries no subtitles), AudioSet 527-class sound tags and top comments — 21 tools sharing one on-disk cache behind a cross-process lock. Needs Python 3.10+, ffmpeg and the vidlens package (pip install vidlens-mcp); a missing install comes back with the two ways to fix it instead of a bare ImportError.

All-media (video/audio/image) sibling of liustack/modlens: turns local media into structured evidence (timeline / transcript / layout regions) injected into the conversation so text-only models can cite concrete frames, seconds and speakers — Evidence, not imagination. One failover chain per media, recorded hop-by-hop in meta.attempts (a fallback is never silent): host-registered multimodal models first, then a configured OpenAI-compatible endpoint (comma-separated multi-key rotation), then local ffmpeg frame-sampling / a local ASR CLI, degrading to metadata-only as a last resort. Absorbs modlens battle-tested rules — cooldown keyed only on auth/quota failures and softened to chain-tail (never dropped), redact-before-clip error text, magic-byte sniff that never trusts file extensions. Optional vision-bridge registers a media-lens wrapper into the host model selector so a pure-text model gains image/audio/video sight. Registering / authenticating multimodal models and true inline audio-video upload lean on two companion plugins, declared as optional peers (absence auto-degrades, never errors): dsh-thinking-levels fronts the model config / registration surface, dsh-llm-openai-completions takes over the OpenAI-compatible transport that serializes file blocks into video_url / input_audio. Privacy: media goes to remote providers by default; strictLocalOnly keeps content on-machine. Prerequisites not shipped with the package: ffmpeg/ffprobe on PATH (frame sampling / reencode / duration probe; degrades to metadata when missing) and a user-configured local ASR CLI for the audio fallback. Hosts 0.1.7 / 0.2.0 / 0.2.1 (6 rc, exact engines.dsh enum). npm: dsh-media-lens.

Sidebar panel for Agnes image and video generation that also ships a node canvas and a short-drama mode. Canvas nodes are chained by extracting the previous shot's tail frame with ffmpeg and submitting it as the next shot's first frame; short-drama mode turns an imported script into storyboard shots. Generation and polling run in the Host process, so tasks survive a page reload or a closed panel, and results plus an 80-entry history are stored on disk.

Codex Appshots for DSH: capture the frontmost active window via global shortcut and seamlessly mount it into the composer for agent queries.

Configure multiple image providers in Settings and call image_generate with the one selected model; images save under generate/image and show inline in the conversation.

Local OCR fallback for text-only routes: when the session model declares it cannot accept images, the attached image is cached locally and its path injected so the model can call ocr_image — PP-OCRv5 + ONNX Runtime on CPU, no API key, images never leave the machine. Silent when the model can see images.

Drive local Adobe Photoshop for game screenshots and photos: layered HDR, multi-zone colour/person masks, SUPIR/Real-ESRGAN super-resolution, native 1:1 vision inspection. Requires Windows 10/11 and Adobe Photoshop 2025; the grading tools do not run on macOS or on earlier Photoshop versions.