dsh-llm-multimodal
Multimodal generation and understanding tools for DeepSeek Harness using OpenAI-compatible endpoints.
Install
$ dsh plugin --profile web add xiaokaizhou/dsh-llm-multimodalPlugin Overview & Capabilities
AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.
Key Capabilities
- Text generation via Harness LlmRuntime routing
- Image generation via OpenAI-compatible /images/generations endpoints
- Video generation supporting text-to-video, image-to-video, and keyframes
- Vision understanding (multimodal chat completions) accepting URLs, Data URIs, and local file paths
- TTS and voice synthesis with voice cloning support
- Music and BGM generation via OpenAI-compatible audio speech endpoints
- Automatic output routing into dsh-media-studio workspace assets or /tmp fallback
Useful For
- Generating pictures, animations, and voiceovers directly within DSH chat conversations
- Performing visual question answering, OCR, and multi-image inspection inside AI workflows
- Cloning character voices using audio samples and generating narrative dialogues
- Orchestrating multimedia assets directly into visual canvas nodes with dsh-media-studio
Who It Fits
- DeepSeek Harness users wanting all-in-one multimodal creation tools in chat
- Content creators and game developers producing character art, voice lines, and video clips
- Developers integrating custom OpenAI-compatible vision, video, and audio gateways
Documented Limitations
- Multimodal tools require OpenAI-compatible endpoints (except generate_text which uses LlmRuntime)
- Video generation uses polling with a maximum wait time around 10 minutes
- Automatic provider configuration UI is temporarily suspended; models are discovered via llm-pi-ai settings
- Voice cloning requires providers supporting MiniMax-compatible file upload and voice clone APIs
DSH Compatibility
Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.
Security Signals
Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.
package.json declares 6 runtime, 2 development, 3 peer, and 0 optional dependencies.
package.json declares DSH bundle metadata.
GitHub reports the repository license as MIT.
A root package.json was captured and can be inspected by the registry.
Public GitHub source metadata is available for this registry snapshot.
Source & Registry Notes
Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.
- Source repository
- xiaokaizhou/dsh-llm-multimodal
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- f4d328f · 2026-10-25
This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.
Repository Activity
- GitHub stars
- 1GitHub stars
- Forks
- 0
- Open issues
- 0
- Last commit
- 2026-09-21
- Last release
- No release detected
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Vision routing and pixel-level tools for text-only DeepSeek Harness agents with built-in free fallback.
View plugin →DeepSeek Harness-native vision toolkit enabling image Q&A, OCR, UI restoration, grounding, and pixel diff for text-only models.
View plugin →Automates invoice retrieval from IMAP mailboxes, OCR extraction, categorization, and Excel summary report generation.
View plugin →A DeepSeek Harness vision plugin that turns pasted images into structured OCR, layout, and semantic evidence for supported text-only models.
View plugin →