Kling
How to Create AI Videos With a Reference Character with Kling O3 Reference
Kling O3 Reference adds character reference images for consistent appearance across clips and supports voice control over individual elements. Sharing the Kling 3.0 backbone (native 4K, 16-bit HDR, Omni Native Audio), it is the right pick when an AI influencer or brand spokesperson needs to deliver lip-synced dialogue across multiple cuts at festival-grade detail. The reference is stronger than Vidu on choreographed tight action; less reference-dense than Vidu Q2 (Vidu accepts 7, Kling O3 Reference reads fewer with stricter ranking).
Step-by-Step Guide
Build a tight character reference set
Kling O3 Reference reads fewer references than Vidu Q2 but applies stricter identity ranking. Build 3-5 high-quality references on Nano Banana 2: front portrait, three-quarter, profile, full-body, and one expressive shot. Quality of each reference matters more than quantity. A blurry or off-angle reference dilutes Kling O3's identity lock.
Use Kling O3 Reference for choreographed-action shots
Where Vidu Q2 wins on dense reference work, Kling O3 Reference wins when the action is tightly choreographed (a specific dance step, a fight sequence, a product gesture that must hit a beat). The Kling motion engine is more disciplined on choreography. For a brand spokesperson hitting a marketing-cue gesture exactly at the second mark, Kling O3 Reference reads tighter than Vidu.
Bake dialogue with Omni Native Audio and per-element voice control
Kling O3 Reference supports voice control over individual elements — useful when the spokesperson speaks in one shot and ambient continues in the next. Specify in the prompt: "First half: character delivers line in English, soft golden lighting. Second half: ambient cafe sound continues, character listens." Lip-sync renders in-pass.
Pick Standard or Pro tier per shot importance
Standard for blocking and social cutdowns; Pro for the festival or broadcast hero shots at native 4K with 16-bit HDR. Render times: Standard 2-3 min, Pro 4-6 min for a 10s clip. The choreography fidelity gap between Standard and Pro is meaningful — for the marketing-cue gesture shot, render Pro.
Layer per-shot prompts in a multi-cut Kling workflow
Because Kling 3.0 supports multi-shot in one pass (up to 6 cuts in 15s), the O3 Reference workflow can deliver an entire dialogue scene with character lock in one render. Specify per-cut prompts inside the same render call; Kling holds the reference identity across all cuts. This is tighter than chaining separate Vidu nodes for each cut.
Save the Kling canvas as a brand-spokesperson template
Once the O3 Reference + character set is dialed in, save the canvas as a brand-spokesperson template. Each new episode reuses the same canvas with new dialogue and locations. Audio bake means each episode ships with synchronized voice and ambience without a separate audio chain.
Prompt Examples
Marketing-cue gesture with native lip-sync. Pro tier renders 4K with the line synced exactly to mouth movement.
Character delivers marketing line "Designed for tomorrow" in English, soft golden hour key light, medium close-up, slight handheld breathing, 5 seconds, native lip-sync, Pro tier 4K
Multi-cut dialogue scene with character lock in one render. Tighter than chaining Vidu nodes per cut.
Multi-cut sequence (12s): 4s wide of character entering office, 4s medium close-up of dialogue line, 4s reverse on listener. Character identity locked across all cuts. Soft daylight throughout. Pro tier 4K.
Choreographed action — Kling O3 Reference's strongest region. The motion discipline is what wins here.
Character performs choreographed gesture: hand rises to forehead in salute, slow turn 90 degrees, soft side rim light, ambient outdoor breeze, 6 seconds, Pro tier
Parameter Tips
Build 3-5 high-quality references on Nano Banana 2 — quality matters more than quantity for Kling O3 Reference.
Use Kling O3 Reference for choreographed tight action; use Vidu Q2 for high-density reference identity work.
Bake dialogue with native lip-sync by writing the line in the prompt — Kling renders mouth movement in-pass.
Multi-cut sequences (up to 6 cuts in 15s) keep character lock tighter than chaining separate single-shot nodes.
Pro tier renders native 4K with 16-bit HDR — use it for the hero shots, Standard for blocking.
Save the canvas as a brand-spokesperson template; reuse with new dialogue per episode.
What to Expect
Kling O3 Reference outputs at native 4K (Pro tier) with synchronized Omni Native Audio in-pass. Render times: Standard 2-3 min, Pro 4-6 min for 10s. Strongest pick for choreographed tight action and dialogue-heavy spokesperson series. Reference density is tighter than Vidu Q2 (3-5 references vs 7) but the motion engine is more disciplined. For dense reference work without dialogue, use Vidu Q2; for budget reference work, use Seedance 2 Omni.
Use Kling O3 Reference on Astorie
Connect Kling O3 Reference with other AI models on Astorie's infinite canvas. No GPU required — start free.
Get Started FreeRelated features
Docs
Related reading
Try Other Models for This Task
Vidu
Vidu Q2 Subject Ref
Vidu Q2 Subject Ref accepts 1-7 character reference images per generation — the densest character-reference slot among the three models in this scenario. For an AI influencer producer keeping "Mia" identical across a 12-week content series, that 7-image character sheet (front, three-quarter, profile, full-body, hands, expression range) gives Vidu more identity vectors than any single-anchor model. The result is the strongest face/jaw/hairline lock across multiple shots, especially when wardrobe and location vary.
View guideByteDance
Seedance 2.0
Seedance 2 Omni adds character reference images to a generation that already accepts up to 12 reference assets — a unique combo of identity lock plus broad multimodal context (audio reference, location reference, palette reference). For an AI influencer producer running high-volume content where each episode varies wardrobe, location, and mood while identity stays anchored, Seedance Omni delivers strong per-clip Sutui economics. It is the pragmatic middle option between Vidu Q2 (densest reference) and Kling O3 Reference (tightest choreography).
View guide