Every claim below reflects shipped, verifiable capabilities as of June 2026.
Capability
V8-MOTION
Pika
Core model
Production system — script to certified master. Cast, storyboard, gate, render, mix, certify, distribute.
Creative clip generator — text/image to short video with stylized effects. Strong at quick social content and visual experiments.
Video engines
✓ Automatic per-shot orchestration across Runway, Kling, Veo, and self-hosted LoRA. Engine assigned by style match with automatic failover. Per-shot provenance records which engine produced each shot.
Pika 2.x model only. Single model, single interface. No multi-engine routing or failover.
Max output length
✓ Episodes up to 10 minutes. Features up to 2 hours. Multi-shot assembly with transitions, grading, and title sequences.
Short clips, typically 3–10 seconds per generation. Extend by iterating, but no built-in multi-shot assembly or episode structure.
Creative effects
Production-grade post: transitions (8 types), color grading (presets + custom), per-act music scoring, pacing controls. Designed for narrative continuity, not per-clip effects.
✓ Distinctive creative effects: Inflate, Melt, Squish, Explode, Crush, Cake-ify, and other stylized transformations. Strong identity for social/viral content.
Audio
✓ 4-layer audio pipeline: dialogue (74 languages via ElevenLabs), per-act music scoring, ambient, SFX. Lip sync verification. Audio mixed and mastered per episode.
✓ Pika has added sound effects generation. Audio is per-clip, not multi-track or scored across a production.
Character consistency
✓ Persistent character bible with reference images, voice profiles, wardrobe notes. LoRA model training. Consistency tracked across all shots and episodes.
✗ No persistent character system. Each generation is independent. Consistency across clips requires manual prompt engineering.
Storyboarding
✓ Scene workshop with 8-position lighting, per-shot camera controls (12 angles, 8 movements), visual timeline with drag-and-drop reordering.
✗ No storyboard. Individual prompt-to-clip workflow. No shot sequencing or visual timeline.
Provenance
✓ C2PA-certified per-shot provenance. Every shot carries tamper-evident origin, engine, operator, and license attestations.
✗ No C2PA. No per-shot provenance chain. No certification for downstream distribution.
Talent licensing
✓ CAST integration pathway — consent and royalty framework for AI likenesses. Per-shot license attestation in C2PA provenance chain.
✗ No talent marketplace. No licensing or royalty framework.
Quality control
✓ 3-gate operator review (budget, creative, final). Prompt hardening engine. Per-shot Claude Vision frame analysis wired in pipeline. Auto-regeneration with pattern learning.
● Regenerate and iterate manually. No structured gate system or automated quality scoring.
API / Agent access
✓ REST API + MCP server with 54 tools. AI agents and humans use the same studio with the same governance.
✓ API available for generation. No MCP server. No governance layer for agent access.
Distribution
✓ Multi-platform distribution hub with per-platform formatting, scheduling, compliance labels, and analytics.
✗ Download only. No distribution pipeline or platform-specific formatting.
Compliance
✓ EU AI Act Article 50, US state compliance modules, jurisdiction-specific disclosure. Configuration-driven.
✗ No jurisdiction-specific compliance tooling.
Pricing model
Per-production pricing. Operator $1,997/mo + $75/episode + $250/feature. Studio $9,997/mo + $60/ep + $150/feature. Price covers the full pipeline: cast, render, audio, grade, certify, distribute.
Credit-metered. Free tier with limited daily credits (watermarked). Standard $8/mo (700 credits). Pro $28/mo (2,100 credits). Unlimited $58/mo (unlimited standard generations). Credits consumed per generation; longer/higher-quality clips cost more credits.