n8n Cloud + self-hosted / Workflow template
Turn your chosen adult character into a coordinated NSFW creator gallery. Make 4, 8 or 12 outfit and scene variants with n8n, reusing the same original portrait for every Qwen Image 2.1 edit.
Cloud uses built-in HTTP nodes. The self-hosted download uses the SpicyAPI community node. Choose the matching setup steps below.
A labelled set of image results for a character gallery, private feed or scene selection. Each result keeps its own scene input and downloaded image in the n8n execution.
How the workflow runs
Reference + scenes
One item per scene
Quote + Qwen edit
Download each image
NSFW / Character / Style
Give a recurring character a complete editorial story: window-light portraits, an evening look and a late-night balcony. Use four scenes to find a direction, then expand the strongest look into a larger set.
What to change first
Reuse the original portrait for each scene. Change the wardrobe and location; keep facial details and hair consistent.
Create a recognizable male character across relaxed portraits, silk-shirt editorials and night-out looks. Keep a few signature traits so the gallery feels like one creator rather than a collection of unrelated faces.
What to change first
Keep the same reference and hair description. Change lighting or wardrobe first, then try a different framing.
Explore a character’s wardrobe without rebuilding the entire scene: a fitted evening look, a themed costume or a different fabric treatment. Reference editing lets you compare the creative change against the original.
What to change first
Describe the garment, color and fabric explicitly. Keep pose and background steady for the first comparison.
Workflow tutorial
Use the Cloud JSON when you want built-in HTTP nodes and no plugin installation. It sends each scene separately, keeps accepted task IDs and downloads the finished images.
Import the Cloud JSON into n8n. In Quote Request, Create Task and Get Task, choose Generic Credential Type → Bearer Auth and select the same credential. Put your API key in that credential’s Bearer Token field, outside the workflow JSON. Leave the workflow inactive while configuring it.
In Settings, fill referenceImageUrl with a reachable HTTPS portrait and set photoCount to 4, 8 or 12. Build Scenes contains the twelve scene presets. Edit those presets to change the shots, outfits or locations. Its shared instruction starts with fashion and complete clothing; adapt that instruction for your own creative direction.
Set a positive maxCostUsd up to 0.10 USD per image. Fill batchId with the current UTC ISO timestamp and leave resumeTaskIdsJson as []. Keep this batchId and the original inputs unchanged when replaying the same batch within 23 hours; use a new timestamp only for a deliberately new batch. Set enabled=true once the reference, key and budget are ready.
Start a manual execution. Each scene gets a fresh quote and a maximum-charge check before submission. Create Task records the accepted ID; Get Task checks every five seconds for at most 60 lookups or 300 seconds per image, plus an in-flight request. Download Image uses no API credential. Successful images reach Batch Results as binary.data, with a temporary resultUrl.
For needs_review results, copy the accepted task IDs into resumeTaskIdsJson as a JSON array, then run with enabled=true. This path only retrieves and downloads those tasks; it skips quoting and creation. If an ID is missing, inspect the saved execution and console first. Do not replay an old batch after the 23-hour guard. Save completed images to your own storage before their URLs expire.
Use this version with the official SpicyAPI community node installed on your self-hosted n8n instance.
Import the JSON into self-hosted n8n with the official SpicyAPI node installed. Select the same SpicyAPI credential in Generate Scene and Get Scene Result. Leave the workflow inactive while setting it up.
Set referenceImageUrl in Settings to a reachable HTTPS portrait. Choose a sharp, unobstructed face. Every scene reuses this original instead of an earlier output.
Open Build Scenes to adapt its twelve named scene entries. The shared instruction keeps the reference character and defaults to complete-clothing fashion examples; adjust that shared wording for your own creative direction. Change outfit, framing or location in each scene entry.
In Settings, set photoCount to 4, 8 or 12 and maxCostUsd to your per-image USD ceiling, up to 0.10 in this starter. The batch ceiling is photoCount × maxCostUsd. Finish the inputs and credentials, then set enabled to true immediately before your first run.
Execute four scenes first. Generate Scene records the accepted task ID; Get Scene Result waits up to 300 seconds and downloads each completed image. Review the results and save the keepers. For needs_review, retrieve the recorded task before another attempt; if no ID exists, inspect the saved execution and console first.
Creative example. Adapt it to your own fictional adult character and source image.
Keep the fictional adult character's face, short auburn hair and green eyes from the reference. Change only the outfit to a cream sweater and the location to a softly lit café. Chest-up portrait, eye-level camera, natural expression.Try this directionReference editing starts with Qwen Image 2.1; new concepts start with Z-Image Turbo. For a more specific NSFW style, explore compatible LoRA models and the Spicy collection before scaling your batch.
The run has one image-edit task per selected scene. Four scenes mean four quotes and up to four submissions; eight and twelve scale the same way. Read the current Qwen editing price, then choose a per-item ceiling that covers your settings. Budget separately for new attempts on rejected creative takes.
Check that every item uses the original portrait, not the previous result. Remove conflicting face or hair details from the scene prompt and reduce simultaneous changes. A seed alone does not lock a face.
Review the item’s settings and current price before changing maxCostUsd. Reduce photoCount if the complete set is larger than your intended budget. This starter accepts a positive ceiling up to 0.10 USD per image.
Telegram /image
Allowed chat check
Quote + reference edit
SendPhoto result
Self-hosted n8n
Private character scene requests and creator previews
One image sent to the allowed chat
Open workflowLoadImage
Describe one edit
Qwen reference edit
SaveImage
ComfyUI
Outfit variants, scene changes and character frames
An edited image saved by SaveImage
Open workflowDownload the template or try the same model in the Playground before connecting the full workflow.