dsh-video-gen
Run the following command in DeepSeek Harness:
dsh plugin install oebeliever/dsh-video-gen
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install oebeliever/dsh-video-gen in your terminal (repo: https://github.com/oebeliever/dsh-video-gen), then restart dsh web to activate the plugin in Settings and chat.
About this plugin
There is no need to leave DeepSeek Harness, open a separate web app, submit a task, and wait for a download. dsh-video-gen connects MiniMax H3 through the easyframe relay so the AI can take you from a prompt to a playable MP4 entirely inside the conversation, with an inline preview card for playback and download.
Five generation modes are covered: text-to-video, first-frame image-to-video, first-and-last-frame transition, last-frame continuation, and reference generation with up to nine reference images and three reference audio clips; setting mode to auto lets the system infer the best mode from your assets. Tasks are automatically polled until completion, and you can check status, cancel, or inspect the queue at any time. A visual workbench above the chat input offers a four-step, Jimeng-style flow — configure visuals, pick a mode and drag in assets, auto-generate a structured prompt, then watch progress and play the finished clip — plus a history library and a quota panel. Model tier, resolution, nine aspect ratios, duration from 5 to 15 seconds, and AI prompt enhancement are all adjustable in Settings.
Built for short-video creators, content planners, and developers who want to fold video generation into their AI workflow on the DSH web profile. The plugin has zero third-party runtime dependencies, runs on Node 20 or above, and keeps your API key local.
Use Cases
- Generate short videos directly in DSH chat with a text prompt
- Drive AI video generation with reference images and audio for consistent style
- Drag assets into the visual workbench, auto-assemble prompts, and render in one click
Best For
- Short-video creators working in the DSH web profile
- Developers who want video generation embedded in their AI conversation workflow
- Content planners who need quick clips without juggling multiple tools
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.