Sign In

Framotive for Qwen Image 2.1

Download

1 variant available

Archive Other

Framotive_Qwen21.zip

262.07 KB

Verified:

Type
Workflows
Stats

18

Reviews

No reviews yet

Published

Nov 21, 2026

Base Model

Qwen 2.1

Hash
AutoV26CCB9C2DD7

Qwen is licensed under the Qwen RESEARCH LICENSE AGREEMENT, Copyright (c) 2026 Hangzhou Tongyi Laboratory Technology Co., Ltd. All Rights Reserved.

Framotive_Qwen21_LLM.png

Framotive for Qwen Image 2.1

Framotive brings LLM chat, reference-based editing, and Qwen-Image-2.1 generation into one ComfyUI workspace. Discuss an idea, assign roles to your reference images, refine the composition, and generate when you’re happy with the prompt.

Four integrated panels keep the workflow compact: Models & LoRA, LLM / MANUAL, Finishing, and Output, with a large image preview beside your controls.

⚠️ Content Disclaimer

Framotive is a creative tool. You are responsible for your prompts, reference materials, generated content, and how you use, share, or publish it. The author does not review or endorse user-generated content.

Please respect third-party rights and follow the licenses and terms of the models, dependencies, and API providers you use.

💡 My Recommendation

I recommend using a larger model through OpenRouter for complex scenes, iterative edits, and character work. These tasks benefit from an assistant that can follow detailed instructions and keep track of multiple requirements.

You can also connect a local LLM. The experience depends on the model you choose and the memory available on your system.

I recommend using DeepSeek V4.1 Flash 1017. Based on my statistics (1,000 requests in a mixed mode—text-only and text-with-images), each request costs $0.001842857

You can use OpenRouter for free by creating an account and generating an API key. Then, either select a model from the list of free options or enter 'openrouter/free' into the 'model' field

✨ Key Features

  • Chat inside your workflow. Discuss ideas and prepare prompts independently of image generation.

  • Cloud and local assistants. Connect to OpenRouter, LM Studio, or Ollama. Local servers run on the same computer as ComfyUI.

  • LLM and MANUAL modes. Work with an assistant or write your prompts directly. The positive prompt stays editable.

  • Task-specific instructions. The built-in system prompt combines shared rules with instructions for your selected task and inputs.

  • Review before generating. Inspect the prompt and proposed settings. You can turn off automatic application of replies.

  • Custom system prompts. Edit, supplement, or replace the built-in instructions and save named presets.

  • Separate chat and generation. Generate another image from the applied prompt without making another chat request.

🧠 14 Image Tasks

Choose a task in Parameters before sending your request.

  • Generate a new image or compose a scene from references.

  • Edit an existing image.

  • Create typography and layouts with requested text.

  • Generate transparent images, remove backgrounds, or edit RGBA content.

  • Inpaint selected regions or extend the canvas.

  • Work with identity, object, clothing, pose, and style references.

You control the selected task. If a request calls for a different task, switch it in the interface before continuing.

🖼 Images & References

Upload images, drag them into the reference area, paste them into the chat composer with Ctrl+V, or reuse a generated result.

Assign each reference a role, such as face, hair, body build, clothing, pose, style, composition, background, or source canvas. Add individual notes, reorder references, and choose which images Qwen uses. The workflow supports up to 16 encoder image slots, including a separately routed reference mask.

The crop editor lets you isolate details and restore the original later. Workflow JSON saves references, roles, notes, and chat history. New chat clears the working context, with an undo option during the session.

For image analysis, select a vision-capable LLM and enable Send images.

🎭 Masks, Transparency & Structural Control

  • Edit selected regions using reference masks, latent noise masks, or optional Union inpainting.

  • Paint white to edit and black to protect, or use the mask’s alpha channel to target transparent areas.

  • Paint annotations and expand the canvas in the image editor.

  • Use PNG alpha for transparent generation, background extraction, and transparent edits.

  • Use optional Union ControlNet with Canny, depth, grayscale, HED, line art, MLSD, pose, and scribble maps.

You can supply prepared maps or preprocess a photo. Install the compatible preprocessors and Union weights for the branches you use.

➕ Positive & Negative Prompts

The Negative OFF / ON switch works in both LLM and MANUAL modes. In LLM mode, the assistant prepares a coordinated positive and negative prompt pair. In MANUAL mode, you can edit both fields directly. Switching Negative OFF preserves the draft and restores your previous CFG and step settings.

At CFG 1, standard sampling ignores negative guidance. The Turbo LoRA negative-mode setting at CFG 1.1 is experimental.

📚 LoRA Library

  • Search and filter compatible local LoRAs.

  • Save descriptions and exact trigger words for reuse.

  • Share information about selected LoRAs with the assistant.

  • Adjust each LoRA’s strength and order.

  • Refresh the catalogue after adding files.

📐 Canvas & Generation Controls

Choose an aspect-ratio preset, enter custom dimensions, or swap width and height. Lock the canvas to have the assistant adapt the composition to your chosen size. A separate permission controls whether the LLM may change ControlNet settings.

Choose Quality, Turbo LoRA, full Turbo checkpoint, or Custom settings. Sampling controls include steps, CFG, sampler, scheduler, denoise, batch size, and seed. The public workflow opens in MANUAL with Turbo LoRA at 1200 × 1200.

⚙️ Optional Processing

  • A second sampling pass.

  • Face refinement with a separate prompt and sampling controls.

  • Model-based neural upscaling.

  • Tiled VAE decoding and prefix KV cache options.

  • Compatible GGUF loading through ComfyUI-GGUF.

  • Local prompt enhancement with Qwen PE-T2I and PE-I2I models.

Optional branches require their corresponding weights and dependencies. The public workflow starts with optional processing off.

💾 Preview & Saving

Preview results in Output and reuse them as references. Save PNGs to ComfyUI’s output directory or another existing folder on the backend computer. Choose a filename prefix and whether to embed metadata.

With metadata enabled, PNGs store the effective generation prompt and settings. Chat history stays in the workflow JSON you save from the editor.

The native folder picker supports Linux with Zenity. On other setups, use the destination path field. API credentials stay in browser storage, separate from workflow JSON.

🌐 Interface & Prompt Language

Framotive uses an English interface. The assistant handles conversation, generation prompts, and lettering inside the image separately. Generation prompts default to English, while requested lettering can retain its original language and script.

💡 Before You Generate

Review complex reference transfers, pose instructions, masks, and proposed settings before generating. The assistant can misread a reference or introduce unwanted details. Check the generated result for identity, anatomy, geometry, and preservation of the areas you intended to keep.

📦 Installation

The ZIP contains the four-node workflow, the custom node pack, a README, and separate Windows and Linux installation guides. Download model weights separately.