AI Video Showreel · 2026

AI Video Showreel 2026

Digital production has a lot of variables. I limit them by giving image and video generation models a tight set of constraints, built in three stages, so the director’s intention survives all the way to the final shot.

The Challenge

Generative models are fast but unpredictable. Left on their own, they choose the framing, the acting and the look, and the director loses the shot.

The Solution

Each stage hands the next one something concrete. A 3D blockout fixes the set, the camera and where the characters stand. Painted frames on top of it fix the look. Only then does the video model animate a shot that is already decided. The workflow is AI-assisted from start to finish, then edited, mixed and finished in DaVinci Resolve. Next I’m extending the 3D side: motion capture from plain video into Blender with accessible tools, so the animation and the expressions carry the emotion a scene needs, and so I can work with actors directly to bring characters to life.

  • Three stages: 3D blockout, painted key frames, then the generated shot
  • Staircase scene for an animated comedy in development, shown from blockout to final, with lip-synced synthetic dialogue
  • Familias Sin Fin films generated with Seedance; ASR and AXS ads animated with Google Veo from Nano Banana key frames
  • Footsteps, room tone and exterior ambience layered per scene
  • End card animated natively in DaVinci Resolve Fusion
  • Credits: music “Wider” by Skygaze (Artlist); synthetic voices made with Qwen3-TTS, Kyutai Pocket TTS (VCTK p360 voice, CC BY 4.0) and Piper

The Result

Most of my portfolio comes from years as a videographer and editor. This reel is where I’m taking that work now, as a digital creator. With the right budget, the same pipeline scales from a single scene to a full series.