VegaNec Candid Presence is a positive-direction Krea 2 concept LoRA built for one stubborn failure: people who look arranged for a catalogue when the scene is supposed to feel lived in.

It pushes prompted fictional adults toward believable support, asymmetry, occupied attention, grounded movement, practical contact, and body language that belongs to the scene rather than the camera.

The model supports standing weight shifts and natural leans; grounded walking and mid-step movement; relaxed seated posture with readable support; expressions and attention directed beyond the camera; practical actions with simple props and surfaces; and wide environmental compositions with believable human scale.

The immutable SFW training package contains 71 images and 71 positive captions across 53 fictional adult identities. Its six behavior lanes contain 15 standing/leaning, 16 walking, 12 seated, 12 candid-expression, 12 practical-action, and 4 wide-environment examples. No identity has more than two training images.

Trigger: veganec_candid_presence

Recommended strength: 0.8

Useful range: 0.8 tested and approved for launch; other strengths were compared but are not claimed as launch recommendations.

Prompting guidance:

Describe the fictional adult, wardrobe, environment, framing, and action you actually want. Let the adapter contribute the shared presence behavior. For prop scenes, specify exactly one simple object, its owner, its state, and the intended contact. For wide scenes, state the subject's action and scale inside the environment. Avoid stacking many pose adjectives before testing the adapter.

Honest limits:

This is a concept LoRA, not a recurring-character LoRA. It cannot guarantee perfect hands, anatomy, props, or composition. It does not replace pose-control, ControlNet, or exact spatial tools. Multi-person scenes and character-LoRA stacking are unverified. Negative weights and monotonic reverse control are unverified. This version is not a slider. Strength 0.8 is the only approved launch recommendation.

Two of twenty-four matched marketing comparisons produced a recurring lower-right duplicated inset/contact-card artifact. Those images were permanently excluded from the public gallery. Reject and reroll any similar collage, inset, or picture-in-picture artifact.

All subjects in the training and launch evidence are fictional adults. The resource is not based on an actual person and is not intended for minor-focused content.