Walkthrough
From Brief to Film — The Agentic Pipeline
Type one sentence. Get a cinematic film. Here's exactly how Twin Studio's agentic pipeline works — with a real example.
The example brief
"Luxury fashion film for a red lehenga, shot during monsoon in Mumbai. It should feel sensual, editorial and cinematic. The model walks through rain-soaked streets at golden hour."
Step 1: Twin o1 parses intent
Twin o1 receives the brief and identifies: fashion film, luxury tone, monsoon Mumbai setting, sensual mood, editorial style, golden hour lighting, rain environment, walking motion.
Step 2: Treatment generation
The harness generates a treatment: "A 30-second fashion film in the style of Sanjay Leela Bhansali's visual richness meets Wong Kar-wai's rain aesthetic. Shot on Kodak Portra 400, golden hour backlight, shallow depth of field. The lehenga's red pops against the grey monsoon streets."
Step 3: Screenplay
Scene-by-scene screenplay: 6 shots, each 5 seconds. Opening establishing shot of Mumbai streets in rain. Model emerges from a doorway. Close-up of lehenga fabric catching rain. Walking shot with fabric flowing. Portrait close-up with golden hour backlight. Final wide shot as model walks away into the rain.
Step 4: Shot list with model routing
| Shot | Description | Model |
|---|---|---|
| 1 | Establishing: Mumbai rain streets | Hailuo 03 |
| 2 | Model emerges from doorway | VEO 3.1 |
| 3 | Close-up: lehenga in rain | Seedance Pro |
| 4 | Walking, fabric flowing | Kling 2.6 |
| 5 | Portrait, golden hour | Atom X → VEO 3.1 |
| 6 | Wide: walks into rain | VEO 3.1 |
Step 5-6: Image → Video generation
For each shot, the harness first generates a hero frame using Atom X (if character consistency is needed) or FLUX Pro. Then the hero frame is animated into a 5-second video clip using the routed model. The character's face stays locked across all 6 shots.
Step 7: Audio + lipsync
MusicGen creates a background score: "cinematic Bollywood string arrangement, building tension, 90 BPM, sensual mood." Audio Z TTS generates any voiceover. If the model speaks, Lipsync X3 syncs the lips to the audio track.
Step 8: Export
The 6 clips are assembled in sequence, audio is mixed, and the final 30-second film is exported as an MP4. Total cost: approximately 300-400 Creative Fuel. Total time: 15-20 minutes.
Then: post directly
Twin Studio posts the finished film directly to Instagram Reels, Facebook, and LinkedIn — with an AI-written caption and trending hashtags. No manual download, no separate upload, no scheduling tool.