AI on the Lot 2026 - QwenVL - Prompt Enhancer
The fourth workflow released for AI on the Lot 2026, made available for ComfyUI 101 workshop attendees to open and follow along if they choose. A prompt enhancer powered by the Qwen VL LLM node: enter a simple prompt and get back a richer, more detailed version ready to use.
概要
AI on the Lot 2026 - QwenVL - Prompt Enhancer is a lightweight ComfyUI workflow that turns a short idea into a rich, production-ready prompt. It uses the AILab_QwenVL_PromptEnhancer node to pass your input through a Qwen-VL language model, which expands key details like subject attributes, mood, composition, and technical cues. The output is displayed with PreviewAny so you can quickly copy it into any downstream image or video generation pipeline.
入力するもの
- ワークフローを読み込み中
ComfyUI で実行できます。 このワークフローはこのページでは実行できません。すべてが読み込まれた状態でComfy Cloudで開くか、ご自身のComfyUI用にダウンロードしてください。
Cloudで試すよくある質問
Qwen-VL is a vision-language large language model. In this workflow it’s used in text-only mode via the AILab_QwenVL_PromptEnhancer node to expand concise ideas into concrete visual descriptions (subjects, style, lighting, composition) that are ready for image/video generation.
It depends on how your Qwen-VL backend is set up. Some installations run locally (no internet required), while others call a remote service (API key required). Open the AILab_QwenVL_PromptEnhancer node and follow its configuration notes for your environment.
State constraints in your input (e.g., “no text on image, no watermark, natural colors, documentary style”). If the node exposes creativity/temperature settings in your build, lower values make results more literal, higher values more creative. You can also iterate by rephrasing your request with explicit dos and don’ts.
Yes. The output is plain text, so you can paste it into the positive prompt field of Stable Diffusion or other text-to-image/video nodes. If your pipeline uses a separate negative prompt, include exclusions in that field or in your initial instruction to the enhancer.