Three video models in production. Each handles a different class of output. The routing layer decides which model touches which brief — no single model handles everything.
In production — Seedance 2.0 · Kling 3.0 · Grok
Real output, not a placeholder
ByteDance (direct partnership)
720p / 5s = 205 credits. Multi-shot syntax: [00:00-00:03] per segment.
Routing rule
Default first choice for narrative and cinematic content.
Kuaishou
Omni mode for lip-sync. Always end prompts with "then settles" for clean loop.
Routing rule
Talking head, lip-sync, product loops, physical simulation.
xAI
Best used for concept exploration before committing a brief to a slower, higher-fidelity model.
Routing rule
Fashion, lifestyle, and editorial concepts; early-stage creative exploration.
Brief intake
Shot description, character refs, tone, and output spec. Parsed by n8n into structured generation parameters.
Model selection
Routing layer assigns model based on task class — narrative, lip-sync, environment, or loop.
Prompt construction
Prompt DSL applied: shot language, timing brackets for Seedance, anchor mechanics for Kling.
Generation + retry
API call with retry logic and fallback routing if the primary model hits rate limits or quality threshold.
QC gate
Duration check, resolution verification, output scoring. Auto-reject and re-queue if thresholds not met.
CDN delivery
Cloudflare R2 upload, URL registered in Supabase, notification to downstream consumer.
Other machines