Multi-Modal Creative Agents for Text-to-Video Storyboarding with CrewAI
This challenge tasks you with building a multi-modal creative agent team. You will use CrewAI to orchestrate specialized agents powered by Gemini 3 and OpenAI GPT 5.1 Pro to generate, storyboard, and critique concepts for short videos from a given text prompt. The system must leverage hybrid instant/deep reasoning, RAG for creative inspiration, and dynamically adapt its thinking budget based on the complexity of the creative brief. The final output should be a detailed storyboard plan, including visual descriptions, suggested camera angles, and a critical evaluation of the concept's potential.
What you are building
The core problem, expected build, and operating context for this challenge.
This challenge tasks you with building a multi-modal creative agent team. You will use CrewAI to orchestrate specialized agents powered by Gemini 3 and OpenAI GPT 5.1 Pro to generate, storyboard, and critique concepts for short videos from a given text prompt. The system must leverage hybrid instant/deep reasoning, RAG for creative inspiration, and dynamically adapt its thinking budget based on the complexity of the creative brief. The final output should be a detailed storyboard plan, including visual descriptions, suggested camera angles, and a critical evaluation of the concept's potential.
Shared data for this challenge
Review public datasets and any private uploads tied to your build.
What you should walk away with
Master CrewAI for defining and orchestrating role-based agent teams with specific tools, goals, and backstories for creative tasks.
Implement hybrid reasoning patterns, utilizing Gemini 3's multi-modal capabilities for initial concept generation (instant reasoning) and deeper visual analysis (deep reasoning).
Leverage OpenAI GPT 5.1 for sophisticated textual critique and refinement of generated storyboards, acting as a 'Creative Director' agent.
Integrate advanced RAG techniques to provide agents with a rich context of film theory, visual styles, and creative brief examples, enhancing creative output quality.
Develop a 'Visualizer Agent' and a 'Critique Agent' capable of processing and generating multi-modal descriptions, ensuring alignment with the text-to-video paradigm.
Design mechanisms for adaptive thinking budgets, allowing agents to allocate more computational resources for complex creative challenges or critical evaluation phases.
Implement a feedback loop within the CrewAI agents, where critique agents can refine storyboards generated by conceptual agents, iterating towards a higher quality output.
How this agent runs
Evaluation will focus on the completeness, creativity, and coherence of the generated storyboard, as well as the quality of the critical evaluation provided by the agent team.
Challenge input
{ "prompt": "String", "genre": "String", "target_audience": "String" }
Agent execution
The configured agent processes the input under the challenge policy.
Evaluated output
{ "storyboard_title": "String", "summary": "String", "scenes": [{"scene_number": "Int", "visual_description": "String", "action_description": "Stri...
- Verify that the generated output includes a title, summary, scenes (min 3), and a critique report.
- Assess if scene descriptions include visual, action, and camera suggestions, demonstrating multi-modal unde...
- Creative Coherence Score target: 0.8
- Python execution harness
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- The evaluation module defines the checks.
Recipe state
This is a preview. The configuration can change before the evaluation recipe is locked.
Run this agent on your dataset and AI stack
Bring your dataset, model providers, and success criteria. We will scope the right managed run for your team.
Scope a managed run[ok] Wrote CHALLENGE.md
[ok] Wrote .versalist.json
[ok] Wrote eval/examples.json
Requires VERSALIST_API_KEY. Works with any MCP-aware editor.
DocsFind another challenge
Jump to a random challenge when you want a fresh benchmark or a different problem space.