<h2 id="-u1leeq1jy"></h2>In this workflow, I’ve built an intelligent ComfyUI setup that automatically improves user prompts to better suit the Flux1 Kontext-Dev editing system — a cutting-edge tool for image-to-image editing .📘 Reference: <a target="_new" rel="ugc" href="https://docs.comfy.org/tutorials/flux/flux-1-kontext-dev">Flux1 Kontext-Dev Official Guide</a><hr /><h3 id="goal-4d9xoq4nn">🎯 Goal</h3>Flux1 Kontext-Dev relies heavily on clear, rich, and well-structured prompts to guide the editing process. However, many users provide short or vague prompts, leading to poor results.This workflow solves that by integrating a local large language model (LLM) using Ollama, which rewrites simple prompts into descriptive, detailed prompts tailored for effective image editing.<hr /><h3 id="how-the-workflow-works-zha9abgg0">⚙️ How the Workflow Works</h3><ol><li>User Inputs:<ul><li>An image for editing.</li><li>A simple or vague text prompt describing the desired change.</li></ul></li><li>Ollama Integration (LLM for Prompt Enhancement):<ul><li>The prompt is passed to <a target="_new" rel="ugc" href="https://ollama.com/library/gemma3">Gemma-3</a>, a vision-enabled LLM running locally via Ollama.</li><li>The model rewrites the prompt into a more expressive and visually descriptive version.</li></ul></li><li>Enhanced Prompt → Flux1:<ul><li>The improved prompt is fed into the Flux1 Kontext-Dev nodes along with the input image.</li><li>Flux1 then performs context-aware image editing based on this high-quality prompt.</li></ul></li></ol><hr /><h3 id="requirements-pz3849ztk">📦 Requirements</h3>To run this workflow, you need the following components:✅ 1. OllamaA powerful local runtime for LLMs and vision models.🔗 Download and install Ollama: <a target="_blank" rel="ugc" href="https://ollama.com/download">https://ollama.com/download</a>✅ 2. Vision Model: <code>gemma3</code>Use a multimodal (vision + language) version of Gemma 3 depending on your system’s VRAM:<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8a88bb3d-b677-40ab-ab77-9deea1fc9550/width=525/8a88bb3d-b677-40ab-ab77-9deea1fc9550.jpeg" />👉 Model Page: <a target="_blank" rel="ugc" href="https://ollama.com/library/gemma3">https://ollama.com/library/gemma3</a><pre><code>ollama run gemma3</code></pre>🔥Uncensored Model: <a target="_blank" rel="ugc" href="https://ollama.com/huihui_ai/gemma3-abliterated">https://ollama.com/huihui_ai/gemma3-abliterated</a><pre><code>ollama run huihui_ai/gemma3-abliterated</code></pre>⚠️ Make sure you're using the multimodal (vision) variant of Gemma 3 to ensure it can process image-based prompts in ComfyUI.<hr /><h3 id="key-benefits-h9fktbk2t">✅ Key Benefits</h3><ul><li>Improved editing accuracy from even simple input prompts.</li><li>Local-first, privacy-safe setup using Ollama and ComfyUI.</li><li>Flexible model choices depending on your hardware.</li></ul><hr /><h3 id="example-6atnqov3x">💡 Example</h3>Input prompt:"change the style to realistic"Enhanced prompt via Gemma-3:"Change the image to a photorealistic rendering, with accurate lighting, textures, and details, while preserving the subject’s facial features, pose, and the existing composition."<hr /><h3 id="multilingual-prompt-support-tnjay9id8">🌍 Multilingual Prompt Support</h3>This workflow supports prompts in any language, including Arabic, and automatically translates them into expressive English prompts that Flux1 can interpret.💬 Example:Input (Arabic):"حول الستايل إلى حقيقي"Enhanced Output (English):"Change the image to a photorealistic rendering, with accurate lighting, textures, and details, while preserving the subject’s facial features, pose, and the existing composition."This makes the workflow highly accessible to non-English speakers while still benefiting from professional-grade prompt enhancement. <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/aed7b397-c935-4a21-abb2-0467e1cc291a/width=525/aed7b397-c935-4a21-abb2-0467e1cc291a.jpeg" /><h3 id="workflow-versions-uiwfcsmad">🧩 Workflow Versions</h3>There are two versions of this workflow available:🔹 Basic Version<ul><li>Designed for ease of use.</li><li>Supports 1–2 input images.</li></ul>🔸 Advanced Version<ul><li>Supports up to 4 input images.</li><li>Includes upscaling at the end of the pipeline.</li><li>Built for professional-quality outputs.</li><li>Based on a modified version of this original workflow from Civitai: 👉 <a target="_blank" rel="ugc" href="https://civarchive.com/models/618578?modelVersionId=1956938">https://civarchive.com/models/618578?modelVersionId=1956938</a></li></ul>

Files with SHA256: 845d3ad06fa724bc64cad8ce3bb85428dc50db1e9116c387a4caf8e30f7cd5c1