Turn your ideas into AI images with Qwen Image 2.1 on Medux. Generate from text alone, or add a reference image to guide the result. Add a negative prompt, choose your format, then preview and download your finished image.
Describe the subject, style, lighting, and setting. Optionally upload a reference image.
The latest task appears below. Processing continues on the server if you leave this page.
Content involving minors, non-consensual intimate imagery, or realistic depictions of real people without their consent is not permitted.
No image yet. Generate one to see it here.
Use the service behind this Workshop in your application or AI agent. Explore the API reference and its corresponding MCP tool before starting an integration.
Bring the image-generation service behind this Workshop into your application. The API reference covers authentication, model selection with model_id, prompt and negative prompt fields, reference images through image_file_ids, and asynchronous task results.
Read the Text to Image Turbo API referenceThe API reference's MCP Tool section documents the corresponding medux_text_to_image tool. Review its published arguments and authorization requirements, then connect your agent using the Medux MCP setup guide.
Open the Text to Image MCP toolWrite one clear subject, setting, and visual style. Select Use sample to start from the example.
Optionally upload a PNG, JPG, or WebP reference image. Select square, portrait, or landscape output, and add a negative prompt for unwanted details.
Select Generate image to start a Qwen Image 2.1 task. Follow its status, preview the completed image, and download it. Account Results keeps the task available when you leave the page.
Create scenes, characters, and illustrations from a written description, then refine one detail at a time.
Upload a reference image and describe the new image you want, such as a different style or setting built around the same subject.
Design a portrait or scene as a starting image for your next video, or combine a portrait with speech in Lip Sync.
Start with one subject and a clear pose. Add clothing, framing, and environment.
Choose a coherent visual direction such as a realistic studio portrait, golden-hour photograph, or painted fantasy scene.
List specific unwanted details such as blurred faces, extra fingers, text, or watermarks. Review each result before using it in another workflow.
It is a browser-based text-to-image Workshop using the Qwen Image 2.1 model (qwen-image-21) through the Medux Text to Image Turbo service. Enter a prompt, optionally add a reference image, then generate, preview, and download the result.
Yes. Upload an optional PNG, JPG, or WebP reference image and describe how the new image should use it. Without a reference image, the model generates from your prompt alone.
Choose square at 1024 × 1024, portrait at 768 × 1344, or landscape at 1344 × 768. Square is the default.
Yes. The Text to Image Turbo API reference documents the service behind this Workshop. Set model_id to qwen-image-21 and pass uploaded file IDs in image_file_ids for reference images. The MCP Tool section documents the corresponding medux_text_to_image tool.
Once a task ID is returned, processing continues on the server. Return in the same browser and account to recover your latest submission, or open Account Results to find your running task and completed image.
The service supports zimage-turbo (default), krea2-turbo, moss-image-sft, qwen-image-21, and qwen-image-edit. Select the model with the model_id parameter or in the Playground.