Skip to content
GenMedia Creative Studio

Using GenMedia Creative Studio

GenMedia Creative Studio provides a centralized interface for experimenting with Google Cloud’s generative media models. The application is organized into three primary areas: Foundation Model Workspaces, Creative Workflows, and Studio management tools.

1. Navigating the Studio

Section titled “1. Navigating the Studio”
  • The Library (/library): A centralized repository to browse, filter by media type, inspect C2PA Content Credentials, and download or reuse all your generated media.
  • Settings & Configuration (/config): Manage your application settings, API configurations, and experimental feature flags.

2. Foundation Models (Direct Generation)

Section titled “2. Foundation Models (Direct Generation)”

Interact directly with the underlying generative media models to test prompts, compare model variants, and create single- or interleaved-modality outputs. See the Foundation Model Workspaces Overview for the shared request lifecycle diagram.

  • Image Generation:
    • Gemini Image Generation (Nano Banana): Generate and conversationally edit high-resolution (1K–4K) images with up to 14 reference images and Google Search grounding using Nano Banana 2 (gemini-3.1-flash-image-preview), Nano Banana 2 Lite (gemini-3.1-flash-image-lite-preview), and Nano Banana Pro (gemini-3-pro-image-preview).
    • Imagen (Deprecated): Legacy Imagen 3 and Imagen 4 models are being deprecated across Vertex AI. Use Gemini Image Generation (Nano Banana) instead. See the Imagen models migration guide for the replacement mapping.
  • Multimodal & Video Generation:
    • Gemini Omni: Generate interleaved text, image, and audio turns in a single workspace using gemini-omni-1.1-flash-preview via the Vertex AI Interactions API.
    • Veo Video Generation: Create high-fidelity videos across Text-to-Video, Image-to-Video, First/Last Frame Interpolation, Reference-to-Video (ASSET & STYLE), and Video Extension using Veo 3.1, Veo 3, and Veo 2.
  • Audio, Music & Speech Generation:
    • Lyria Music Generation: Generate instrumental and vocal music tracks from text prompts, custom lyrics, and reference images using Lyria 2 (lyria-002) and Lyria 3 (lyria-3-clip-preview, lyria-3-pro-preview) with C2PA provenance preservation.
    • Gemini TTS: Synthesize expressive single- and multi-speaker speech with natural-language style direction, inline vocal tags, and integrated Praat/Librosa acoustic telemetry (gemini-3.8-flash-tts, gemini-2.5-flash-tts).
    • Chirp 3 HD: Generate high-definition speech across 30+ locales with speaking rate controls and custom IPA / X-SAMPA phonetic pronunciations.
  • Text & Prompt Ideation:
    • Gemini Writers Workshop: Collaborate with Gemini on multimodal prompts, customize system instructions, and save reusable PromptTemplate records to Firestore.

3. Creative Workflows

Section titled “3. Creative Workflows”

Creative Workflows are opinionated, multi-step pipelines that chain Gemini reasoning, Nano Banana image generation, Virtual Try-On, Veo video synthesis, and MoviePy post-processing. See the Creative Workflows Overview for a full comparison matrix.

  • Storyboarder: Expand a story concept and character reference into a multi-scene script with Gemini, synthesize character-consistent keyframes with Nano Banana, animate every scene with Veo, and stitch a master MP4 movie.
  • Shop the Look: Dress a fashion model sequentially across multiple apparel items using Vertex AI Virtual Try-On, verify every step with a Gemini visual critic (select_best_image & final_image_critic), and generate a Veo runway clip.
  • Character Consistency: Extract a structured FacialCompositeProfile from reference photos, synthesize candidate scenes, select the best likeness via Gemini, outpaint to 16:9, and animate with Veo.
  • Interior Design: Convert a 2D floor plan into a 3D isometric visualization, zoom into individual rooms, iteratively restyle finishes in the Design Studio, and assemble a guided Veo walkthrough tour.
  • Virtual Try-On (VTO): Combine real or synthetically generated (VirtualModelGenerator) model portraits with product garments and animate the result with Veo.
  • Starter Pack: Bi-directionally convert between lifestyle outfit photographs and flat-lay “Starter Pack” product moodboard collages.
  • Motion Portraits: Animate still portraits or live webcam selfies using curated cinematography styles (PORTRAIT_STYLES), Gemini scene direction, Veo Image-to-Video, and automatic GIF conversion.

4. Post-Production & Experimental Labs

Section titled “4. Post-Production & Experimental Labs”
  • Pixie Compositor: Stitch multiple Veo video clips with transitions (concat, crossfade, wipe), layer Lyria or TTS audio tracks onto video, and export looping GIFs.
  • Labs & Experimental Workflows: Explore incubating workflows at /labs—including Retro Games, Brand Adherence, Character Asset Sheet, Guideline Analysis, Object Rotation, and Banana Studio.