What's New

Upload up to six product photos to make a 15–180 second video with sound. Choose the format and look; longer videos use separate scenes joined into one MP4.

• Multi-photo scene builder

  Each photo becomes a scene beat—jewelry, skincare, sneakers, and electronics all get cinematic treatment from a single upload.

• Sound and direction included

  The generator adds synchronized audio and camera motion so the output is share-ready without post-production.

The trending finger-frame reveal effect now has a dedicated workflow on Voor. Upload one portrait, pick a style, and get a complete finger-frame video—no editing app required.

• One photo in, one video out

  The generator handles the opening pose, illustrated style reveal, and downward hand release in a single pass.

• Multiple illustrated styles

  Choose from several looks—each applies a different artistic treatment during the reveal sequence.

MiniMax H3 generates video with synchronized audio from text or image prompts. The model is now integrated into the Voor generator with six real examples, exact prompts, and review notes.

• Audio in the first pass

  Dialogue, ambient sound, and music arrive with the video—no separate audio generation step needed.

• Text and image input

  Start from a scene description or guide the shot with a reference image for consistent subjects and setting.

Two new video-model pages collect real examples, focused prompt patterns, and production review guidance. Seedance 2.5 is now fully integrated into the generator for live rendering.

• Seedance 2.5

  Plan longer multimodal story beats with assigned references, continuity locks, camera direction, dialogue, and native sound.

• Muse Video

  Build prompt-faithful social scenes with stable subjects, reference-driven creation, coordinated motion, and audio payoff.

Three newly connected model families now have focused workflows on Voor: Seedream 5 Lite and Pro for reasoning-led images, HappyHorse 1.1 for dialogue and reference video, and Seedance 2.0 Mini for fast creative drafts. The Kling 3.0 Turbo guide is also live, with its current Kling v3 generator labeled separately until the Turbo API is connected.

• Seedream 5 Lite + Pro

  Move from fast composition tests to demanding high-quality image work inside one model family.

• HappyHorse 1.1

  Switch between text, image, and reference-to-video modes for narrative clips with audio direction.

• Seedance 2.0 Mini

  Test hooks, product loops, and motion ideas quickly before promoting the winning brief.

• Honest Kling mapping

  Read the Turbo workflow and examples without mistaking the currently available Kling v3 model for the unavailable Turbo release.

Workflow Agent conversations now save to the server, complete workflow versions stay available in cloud history, and multi-step turns can inspect tool results before continuing. Interrupted generation work can reconnect after a deployment instead of disappearing with a browser tab.

• Cloud conversation history

  Continue the same workflow discussion across sessions and devices.

• Version browser

  Load older workflow versions from the server instead of relying on a short local snapshot list.

• Bounded iteration

  The agent can use tool observations and revise its next step without entering an endless background loop.

• Safer recovery

  Persisted turns and run leases help active jobs reconnect cleanly after app updates.

Two focused image workflows are now easier to find across Apps, Collections, and the top navigation. App Icon Generator uses Recraft v3 for crop-safe visual concepts; AI Image Translator locks GPT Image 2 Edit to replace visible copy while preserving the source layout.

• Platform-ready icon package

  Explore a visual metaphor, review the smallest crop, and export the selected icon into common platform sizes.

• Layout-aware localization

  Translate posters, menus, signs, and product graphics while keeping people, products, spacing, and hierarchy stable.

Pay-as-you-go credit packs are back on the web pricing surface and can be opened from tool pages in a full-screen Stripe Checkout. Use a one-time top-up when a project needs more generations without starting a subscription.

• No subscription required

  Purchase a credit pack for the current project and keep using the same account balance across image, video, audio, and workflow tools.

• Cleaner checkout

  The expanded checkout keeps payment controls visible on smaller screens and returns to the active creation flow after completion.

We shipped a dedicated Happy Horse AI workflow page on Voor AI: jump straight into Seedance 1.5 Pro with Text to Video tooling—built for creators searching for the Alibaba-class model that topped the Artificial Analysis Video Arena.

• From prompt or reference

  Start with a scene brief or guide motion from an image while the generator stays on the production video stack.

• 1080p storytelling workflow

  Iterate on pacing, shots, and synced audio directions in one place instead of juggling disconnected tools.

OpenAI's newest native image model, GPT Image-2 (the engine behind ChatGPT Images 2.0), is now available on Voor AI in both text-to-image and image-to-image flows. Pick it from the model list and get sharper text, smarter layout, and multilingual typography out of the box.

• Text & layout that holds up

  Long headlines, multi-word labels, and non-Latin scripts stay legible—posters, slides, maps, and infographics come out review-ready.

• Thinks before it draws

  Reasoning and tool use are built in: the model can parse a reference, check world knowledge, and plan a grid before generating pixels.

• Text-to-image & image-to-image

  One model for both flows. Start from a prompt or drop a reference photo in, and keep identity, wardrobe, and brand cues consistent across edits.

• Flexible aspect ratios

  Horizontal, square, or vertical frames spanning 3:1 to 1:3, with API resolutions up to 4K in beta—one prompt seeds a full asset set.

3D model generation turns prompts (and compatible inputs) into meshes you can inspect in the product preview—then export for rendering, games, or print-ready workflows. Starting with a flat logo, illustration, or photo instead? The 2D to 3D converter returns a downloadable GLB from one image.

• Scene-style preview

  Orbit the generated mesh like a viewport pass before you commit to an export.

• Mesh export

  Download geometry for DCC tools, real-time projects, or print prep—whatever your pipeline needs next.

The next-generation Seedance 2.0 experience: stronger motion, multimodal references, and audio-aware storytelling—served through the production Seedance 1.5 Pro model in the generator.

• Text & image to video

  Start from a prompt or guide the shot with references for more consistent characters, wardrobe, and camera grammar.

• 1080p with synced sound

  Pair visuals with dialogue, music, and effects in one pass on the Seedance stack.

Nano Banana 2 brings Gemini-class image generation with Flash-speed turns, sharper on-image text, and rich detail for ads, social, and product visuals.

• 1K, 2K, and 4K output

  Pick the resolution that fits your channel, from mobile stories to large hero art.

• Edit & iterate

  Refine prompts, adjust style, and keep brand-consistent looks across a series of generations.

FLUX2 klein is the distilled FLUX2 stack for Text to Video—built for fast rerolls, crisp subject detail, and low-latency creative sessions.

• Rapid iteration

  Explore many motion directions before you commit to a longer render on heavier models.

Vidu Q3 turns stills or prompts into short clips with fluid motion and audio that stays in sync—ideal for ads, shorts, and product demos.

• Marketing-ready clips

  Go from a key visual to a moving storyboard in one workflow on Voor AI.

We're excited to launch Nano Banana, Google Gemini 2.5 Flash Image Generator—cutting-edge image generation in the browser.

• Text-to-image

  Describe a product, person, or scene and get balanced images with clean lighting and structure.

• Multi-image processing

  Edit and blend multiple references while keeping style and character consistent across a project.

• Fast generation

  Rapid iteration returns options quickly so you can decide faster.