Fact-check note: "From sketch to video and playable 3D"

For: Oleg Sidorkin
Re: app.cinevva.com/blog/2026-03-04-sketch-to-animated-3d-character
Checked against: cinevva-site repo (marketing site + worker)

---

Verified ✓

• Tool names (Image Generator, Video Generator, 3D Model Generator) and link /tools/hunyuan3d match the codebase.
• Author (Oleg Sidorkin, CTO) is correct.
• Video workflow: Start + end frame interpolation, 9:16, 3–15 s duration, and storyboard mode are all supported.
• 3D pipeline: Hunyuan3D and "3D Model Generator" are correct; auto-rigging (Tripo) and "upload GLB → rigged model + animation library" match the implementation.
• Cinevva Engine and "Reflects early 2026" are accurate.

---

Correction needed

"Seedance 1.5 Pro" — In this repo, video is implemented with Kling 3.0 Pro (fal.ai), not Seedance 1.5 Pro:

• worker/src/genai/fal-video.ts uses fal-ai/kling-video/v3/pro/...
• worker/src/index.ts tracks cost as model: 'kling-3.0-pro'
• tools/video.md says: "Generate short AI videos from text or images using Kling 3.0 Pro"

The API/product name is "seedance" (e.g. /genai/seedance/generate), but the underlying model is Kling 3.0 Pro. Seedance 1.5 Pro is a different model (ByteDance). Public sources often cap Seedance 1.5 Pro at ~10 s; 15 s is supported for Kling here.

Recommendation: In the article, name the video model "Kling 3.0 Pro" (or e.g. "Video Generator (Kling 3.0 Pro)") so it matches the marketing site and docs. If the app backend actually uses Seedance 1.5 Pro, align the app copy and tools/video.md so they describe the same model.

---

Not verifiable in repo

Cubcoats, Kali the Kitty, Mimi Chao, and related business claims are external; not checked.