huashenglian/dsh-omni-workstation
Any-to-any workstation plugin for DSH: multi-card VLM failover, a local vision toolkit, image generation including ComfyUI, async video generation, TTS and voice cloning.
One plugin for every modality the harness can touch. Image analysis runs through an ordered multi-card VLM failover chain, backed by a six-tool local vision toolkit; image generation supports ComfyUI workflows; video generation runs asynchronously across several cards with an AI tool builder; and TTS plus voice cloning cover three cloud and four local providers. All of it is configured from a single settings page that saves as you edit, rather than one config file per provider.
Install
dsh plugin --profile web add github:huashenglian/dsh-omni-workstationThe GitHub form from the README's install block (which also lists a local-directory form and a packed tarball); the repository is the source of truth and carries its own cordis.patch.yml, so dsh plugin add appends the bundle to dsh.profile.bundles automatically.
Compatibility
A self-activating bundle for the web profile — no manual patch editing. Image generation can drive ComfyUI workflows, so that path needs a reachable ComfyUI instance; provider credentials for the cloud VLM, video and TTS backends are configured on one auto-saving settings page (English and Chinese).
Details
- Repo: huashenglian/dsh-omni-workstation
- Category: Vision & Multimodal
- Stars: 5
- Version: 0.1.2 (README badge); MIT
Alternatives
lijian-ui/dsh-vision-toggle · NeoHotaru/dsh-read-animation · niuniuaba/dsh-subagent-vision