Characters
Train your own character (LoRA)
Show a few photos, tendre.red learns the face and draws it anywhere.
Download tendre.red
LoRA training teaches the AI a specific person or character from a handful of images, entirely on your own machine. Once trained, you can place that same character in any scene, outfit or pose, with a consistent identity across every render. The photos never leave your GPU.
- Learns a face from a few reference photos, trained locally
- Consistent character across unlimited new images
- Stack multiple LoRA (character, scene, style, action)
- Reusable in generation, chat and video
From a few photos to an infinite character
A LoRA is a small model that teaches the AI what one specific person or character looks like. You give tendre.AI a handful of reference images, it studies them on your own machine, and it distills the essentials: the face, the hair, the overall look. Training runs on your GPU, so the source photos never leave your computer.
Reuse it everywhere
Once trained, the character is yours to place anywhere: new outfits, new locations, new poses, always recognizable. You can stack several LoRA at once (a character, a scene, a style, an action), each with its own strength, and reuse the same character across image generation, the chat companion and video. It is the difference between a random face every time and a consistent cast you actually own.
Characters
Meet your AI girlfriend tonight.
Download tendre.red, activate once, and create your companion in private, on your own GPU. No subscription.
No NVIDIA GPU? Start 100% in the cloud: 10 free images every day, then 1 credit per image. An NVIDIA card (8GB+) unlocks unlimited local generation.