TVVV · tvvv.ai
Changelog
New features, models and pricing changes, newest first.
v1.12
- New: World Model — walk through a real-time AI world started from a sentence or an image (Lingbot World 2, Abot World 0); WASD to move, arrows to look. Plus MiniMax H3 Director: direct the next shot in text while it plays, with sound
- Music adds My Singers: record or upload 15–30s of a cappella vocals to clone your own singer and use it on any song
- New: Live Sketch — draw or paste on the left, AI repaints it on the right in ~1–2 s, with a follow-the-drawing / AI-freedom slider
- Virtual Model adds dedicated try-on engines: Bria Pro try-on (up to 3 items, ~80% cheaper) and Google single-item; GPT Image 2 stays as Freestyle
- Canvas agent: selected nodes show in the composer, so you can say “make this night time” or “turn it into a video”
- Seedance pricing: Ultra at our cost, Pro+ +5%, Pro +8%, pay-as-you-go +10%; other models: pay-as-you-go +40%, Pro +30%, Pro+ +20%, Ultra +15%
- New changelog page (linked in the footer)
v1.11
- New tools: brush inpainting, video dubbing / lip sync / auto captions, voice cloning, multi-version ads
- New models: Vidu Q4, FLUX.3 Image Klein, Krea K3 Edit, Gemini Omni Flash v1.1
- Default image model is now Nano Banana 2.1
- Character library in image, video and canvas references (public and mine); characters now need a name
- Homepage: 5 hero banners, one-row characters with author; three-column image / video menus
- Canvas image regeneration no longer reuses the previous picture
v1.10
- Nano Banana 2.1 and FLUX.3 Image launched; MiniMax H3 Max style videos
- Fixed Seedance reference roles; copying a work restores its prompt and reference images
- Official example thumbnails for Fun and Video Styles; English badges in the model picker
v1.9
- Image upscaling (SeedVR)
- Agent can run on GPT-6 Astra
v1.8
- Seedance 2.5 Preview: 480p draft first, then render the final
- Lyric music videos, MiniMax Hailuo 3 Max
- Video upscaling (Topaz)
v1.7
- Motion transfer (DreamActor v2) and character replacement (Wan Animate)
- Free voiceover model
- Stem separation (vocals / music)
v1.6
- Seedance 2.5 full release (up to 30 s, 1080p)
- One-click background removal
- GPT Image 2.5
v1.5
- Image to 3D (Hunyuan 3D, TRELLIS 2, Tripo)
- Mureka songs with lyrics
- Agent with Claude / GPT Astra text models
- MiniMax Hailuo 3
v1.4
- fal.ai added: more video, image and music models, routed to the cheaper provider
- AI music: Lyria 3, ElevenLabs, MiniMax Music, ACE-Step, Stable Audio
- Ideogram v4, Hailuo 3 Max Turbo
- Membership credits adjusted (new purchases and renewals only)
v1.3
- Vercel AI Gateway route: Kling 3.0, Seedance 2.0 Fast, Wan 2.6 / 2.7, Veo 3.1 Lite and more
- More video models billed per second; GPT Image billed by usage
- Super creators billed at cost
v1.2
- Credits are reserved on submit and charged only on success; failures are refunded
- Whole-credit pricing
v1.1
- Memberships (Pro / Pro+ / Ultra) and company plan
- Seedance 2.0 / 2.0 Mini
- Terms, privacy policy and work retention
- Claude and GPT Astra text models
v1.0
- Big video update: Seedance 2.5, Seedance 1.5 Pro, Wan 3.0, Grok Imagine Video, Kling 3.0 Motion Control
- Images: Muse Image, Qwen Image 3, Recraft v4.1
- Voice: MAI Voice, Fish Audio S1
- AI avatars
v0.6
- Seedream 5.0 Lite / Pro, Grok Imagine Image 2.0
- Free Kokoro voiceover
- DeepSeek V4 Flash
v0.5
- AI voiceover (Fish Audio S2.1 Pro)
v0.4
- GPT Image 2
- Library: generations are saved automatically
v0.3
- AI agent chat for creating
- Nano Banana (Gemini Flash Image)
v0.2
- Nano Banana Pro (Gemini 3 Pro Image)
v0.1
- tvvv.ai launched: AI video, Seedream 4.5 / FLUX.2 images, DeepSeek chat