Images and task progress shared between chat and the creation page.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-10-24
License
Apache-2.0
01What can it help you accomplish?
Create or edit images inside Harness
Images and task progress shared between chat and the creation page.
Harness users creating images in the panel or with an agent
02How to install into DeepSeek Harness
Prerequisites
- Prepare your own model weights and inference hardware; the plugin includes neither.
Installation steps
- 01
Download the tgz from Releases and run `dsh plugin --profile desktop add /absolute/path/dsh-qwen-studio-<version>.tgz`. On Windows, replace `dsh` with `dsh.cmd` or `cmd /c dsh.cmd`; invoking `dsh` directly fails with EINVAL.
$ dsh plugin --profile desktop add /absolute/path/dsh-qwen-studio-<version>.tgz
- 02
Quit and reopen DSH, then configure the local MLX environment or your NVIDIA gateway in Image settings.
Verify the integration
Not specified by the author
03DSH integration and capability boundaries
Native Qwen Image 2.1 generation and reference editing inside DSH.
Native reference-image editing
Generate an image from a reference image and editing instructions.
04Who is it for? When not to use it?
Good for
- Harness users creating images in the panel or with an agent
Not for
- Model weights are separate; quantized weights start around 40 GB. Low-memory Apple Silicon and 10–12 GB NVIDIA setups have not been verified in this version.
- Native editing currently uses linear scheduling and sizes divisible by 32, and cannot be combined with control images or masks.
- The private gallery category neither hides images in chat nor provides encryption or a password lock.
- Linux remote GPU acceptance is incomplete; successful editing is not guaranteed for every instruction.
05Compatibility, maintenance and safety notes
- Model weights are separate; quantized weights start around 40 GB. Low-memory Apple Silicon and 10–12 GB NVIDIA setups have not been verified in this version.
- Native editing currently uses linear scheduling and sizes divisible by 32, and cannot be combined with control images or masks.
- The private gallery category neither hides images in chat nor provides encryption or a password lock.
- Linux remote GPU acceptance is incomplete; successful editing is not guaranteed for every instruction.
README author claims only.
06Frequently asked questions
How do I install Qwen Studio in DSH?
Download the tgz from Releases and run `dsh plugin --profile desktop add /absolute/path/dsh-qwen-studio-<version>.tgz`. Quit and reopen DSH, then configure the local MLX environment or your NVIDIA gateway in Image settings. On Windows, replace `dsh` with `dsh.cmd` or `cmd /c dsh.cmd`; invoking `dsh` directly fails with EINVAL.
Can a DSH agent edit a reference image?
Generate an image from a reference image and editing instructions.
Does installing the DSH plugin provide model weights?
No. Provide your own weights and inference hardware; quantized weights start around 40 GB.
07Related DSH workflows
modlens
by liustack
Vision bridge for text-only models: paste an image, get structured JSON evidence (OCR, layout, semantics).
agent-vision-toolkit
by anionex
Vision toolkit and skills that give text-only LLMs eyes — multi-image understanding, image Q&A, OCR, frontend UI restoration and GUI automation, with optional agent integration.
dsh-vision-router
by ysr666
Eyes for text-only DeepSeek Harness agents: built-in free vision chain (no key) + pixel-level vision tools (Q&A, grounding, crop, pixel diff, colors, OCR, SVG trace, cutout, screenshots). One-command install, no Python, image turns work like ordinary tool-calling turns.
dsh-vision-toolkit
by anionex
为纯文本 DSH Agent 提供 10 个结构化视觉工具：意图感知图片问答、长截图 OCR、原始像素 grounding、UI 还原、像素 diff 等
08Data and sources
Native Qwen Image 2.1 generation & reference editing for **DeepSeek Harness**.
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-10-24. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.