AI Animated NFC Product Interactions

About the project

Project title 3D-AI Animator for NFC Product Video (hand-tap interaction, hyperrealistic) Description I'm looking for a specialist experienced in AI-assisted generative animation (ComfyUI, Stable Diffusion/FLUX + ControlNet, IP-Adapter, or an equivalent pipeline) to produce 3 short video sequences showing a person tapping a smartphone on physical NFC products (keychains, counter plates, wall plates). I have 3D model files (.GLB) of the products and reference renders to guarantee brand fidelity. The focus of this project is achieving photorealism in the hand-object-phone interaction, without altering the design, logo, or geometry of the real product. Non-negotiable technical requirements Product fidelity: the product in the final video must be identical to the provided GLB/sample (shape, color, texture, branding) — no hallucinated logos, QR codes, or textures. Correct hand anatomy: no artifacts, deformations, or incorrect number of fingers. Lighting coherence: contact shadows, reflections, and depth of field consistent with each scene's environment. Real proportions: size relationship between phone, hand, and NFC device must be anthropometrically correct. Resolution: minimum 2K export, suitable for web and video. Atomic deliverable Each "shot" = 1 scene (product + context + interaction) at a specific angle/framing, 3-4 seconds long, approved against the criteria above. All variants are quoted based on this base unit. The 3 animations to produce 1. Object swap — fixed table Camera and table are fixed. The same hand and phone perform the tap across all 3 variants. What changes between shots: lighting, and the product itself (WhatsApp version → Instagram → Web). The phone screen should display the corresponding app in each version (can be handled as a post-production overlay, not necessarily AI-generated). Duration: 3-4 sec per variant. 2. Counter display with selection All product models are laid out on a counter. A person (torso/arm visible, never full body) appears and taps on one specific item. This is the shot with the most character visibility of the three. 3. Narrative journey (3 moments) A sequence of 3 scenes with WhatsApp-branded products, each with a different background, character, and object (no repeated character between scenes): Keychain on a personal keyring Tap on a counter-display object Tap on a wall plate, while the person is leaving the location (compound action: tap + exit in the same shot) Constant across all 3 scenes: same angle and same proximity/framing to the object, even as background, character, and product change. The person is never shown full-body in any shot. What I'll provide .GLB files for all products Reference renders (multiple angles, neutral background) Real product measurements Specification of critical brand zones (logo, QR code, non-modifiable textures) What I'm looking for Proven experience with ControlNet/IP-Adapter/LoRA on real product imagery Portfolio of prior hand-object interaction work (essential — please include examples) Ability to generate short animation (3-4 sec) with temporal consistency, not just static images Nice to have: compositing experience (Nuke/After Effects) for the phone screen overlay Final deliverables 5 shots total (3 variants for idea 1 + 1 shot for idea 2 + 3 shots for idea 3 — final count to confirm) Format: video, minimum 2K, [add total duration/fps preference] Maximum [N] revision rounds included per shot Budget and timeline [to be defined]

Skills required