Browse categories
Explore
Fiverr Pro
English
$
USD
I will fix build comfyui workflows consistent ai influencer model custom lora training
Donald
About this gig
Struggling with broken ComfyUI workflows or an AI influencer whose face changes between images?
I build and fix ComfyUI workflows and train custom LoRA models for consistent characters, realistic images and repeatable content creation.
Services I offer:
- ComfyUI setup, workflow repair and custom node troubleshooting
- FLUX or SDXL LoRA training for your AI influencer or virtual model
- Dataset review, image selection and captioning
- Character consistency across poses, outfits and backgrounds
- Text-to-image and image-to-image workflows
- ControlNet guidance, inpainting and image upscaling
- Batch generation and reusable prompt settings
- RunPod cloud setup or local installation
- Image-to-video workflow setup, quoted separately
Depending on your package, receive workflow JSON files, trained LoRA files, sample prompts, test images and a step-by-step usage guide. I test the agreed setup and review character consistency against your references
Send your goal, reference images, workflow errors and GPU details before ordering. We will agree on the model, deliverables and scope. Cloud GPU costs and paid tools are separate.
Results depend on your dataset and chosen model.
Delivery style preference
Please inform the freelancer of any preferences or concerns regarding the use of AI tools in the completion and/or delivery of your order.
Get to know Donald
Donald
AI Video Ads, UGC, Product Videos and Social Creatives
- FromUnited Kingdom
- Member sinceApr 2026
- Avg. response time1 hour
Languages
English, French, Spanish, German
I create realistic AI video ads, UGC-style content and product videos designed to grab attention and communicate your offer clearly. I can turn your product, script, images or idea into polished short-form content for TikTok, Instagram Reels, Facebook and YouTube Shorts. From AI avatars and product demonstrations to voiceovers, captions, B-roll, hooks and final editing, I handle the creative process from concept to finished video.
Contact me now to get started
My Portfolio
FAQ
Can you fix my existing ComfyUI workflow?
Yes. Send your workflow JSON, error log and expected output. I will confirm whether it fits the Basic package before you order. Repairs cover an agreed issue in one environment; rebuilding the workflow or adding features requires a separate scope.
Can you train a LoRA for a consistent AI influencer?
Yes. I train one character LoRA on an agreed base model and test it across selected poses, outfits and backgrounds. We agree on references and acceptance examples first. Consistency depends on dataset quality, settings and scene complexity.
Do you work with FLUX, SDXL and Stable Diffusion?
Yes. We choose one supported base model for your goals and hardware. A LoRA is trained for a compatible model family; one training package does not include separate FLUX, SDXL and video LoRAs.
What images do you nede for LoRA training?
Send clear, varied images of the same character with different angles, expressions and lighting. Standard includes reviewing and captioning up to 30 supplied images. Creating a dataset from one reference is a separate service.
Can you install ComfyUI locally or on RunPod?
Yes. I can scope local installation or RunPod cloud setup after checking your GPU, VRAM, operating system and required models. Premium includes one RunPod environment. Cloud rental, storage and paid services are billed separately.
Can you create AI influencer videos and Reels?
Yes, as a custom add-on. I can scope a compatible ComfyUI image-to-video workflow, including WAN where suitable. Video duration, resolution, motion and identity checks are agreed first. Lip sync and video LoRA training are separate.
Can you improve realism, poses and image quality?
Yes. Depending on scope, I can add ControlNet guidance, inpainting, upscaling and prompt adjustments to improve pose control and realism. Send examples of the problem. These features are selected for your workflow, not all included in every package.
Can I use my character for commercial content?
Tell me your intended use before ordering so we can select suitable models and assets. Use original characters or references you have permission to use. Commercial use depends on the licenses and rights attached to the chosen models, images and other assets.
