AZ8 - AI video creation workbench, Agent reads materials and status in real time

What is AZ8

AZ8 is an Agentic Studio (intelligent agent studio) for AI video creationpractitioners, centered around expanding creation and aggregating multiple top-tier video models such as Seedance, Wan, MiniMax, and more. The platform has a built-in creation Agent "Zumi" that can proactively read canvas materials and project status and provide next-step creation suggestions; it also supports 3D precise control, image-to-video, text-to-video, and community work showcases, reconstructing the long-video production workflow for AI video through Agent + canvas.

AZ8's main features

Infinite canvas: Provides a boundless creation space that supports the free arrangement and aggregation of multimodal materials such as text, images, videos, audio, and 3D scenes.
Creation Agent Zumi: Reads canvas materials and project status in real time, proactively providing creators with next-step creation suggestions and decision support.
Multi-model aggregation: Integrates multiple top-tier video generation models such as Seedance, Wan, MiniMax, and Seedream, allowing flexible invocation of different capabilities.
3D precise control: Through methods such as 3D clay models, precisely controls the space and motion of characters and scenes in videos.
AZ8 Theater: Provides a dedicated online showcase, screening, and community exchange space for AI-native short films and creators' works.
Image-to-video/text-to-video: Supports directly generating high-quality AI videos from images or text descriptions, covering multiple creation starting points.
Community collaboration: Creators can share works and gain inspiration within the platform, forming a content ecosystem centered around AI video.

How to use AZ8

Register and log in: Visit the AZ8 official website https://az8.art/zh，完成账号注册，可直接进入创作界面。
Canvas creation: Drag and drop to add materials such as text, images, and video clips on the infinite canvas, freely arranging multimodal content.
Invoke Agent: Click to activate Zumi, which automatically analyzes the existing materials on the canvas and provides next-step creation suggestions or executes actions.
Model generation: After selecting materials, choose models such as Seedance and Wan, then enter prompts to generate or extend video clips.
Precise control: Upload 3D clay models or set keyframes to precisely constrain character actions and camera movement.
Work publishing: After completing creation, submit the work to AZ8 Theater to showcase it in the creator community and gain exposure.
Community interaction: Browse other works in Theater, gain inspiration, and participate in discussions and collaboration.

AZ8's core advantages

Agent-driven creation: With the built-in Zumi intelligent agent, it can understand canvas status in real time and proactively provide creation suggestions, lowering the decision-making threshold for long-video production.
Infinite canvas workflow: With the boundless canvas as the core, it integrates multimodal materials such as text, images, videos, audio, and 3D scenes, reconstructing the video creation process.
Flexible multi-model scheduling: Aggregates top-tier video models such as Seedance, Wan, and MiniMax, invoking different capabilities on demand and avoiding the limitations of a single model.
3D precise control: Supports 3D constraint methods such as clay models, enabling pixel-level precise regulation of character actions and camera movement.
Creator Community Ecosystem: AZ8 Theater provides a platform for showcasing and exchanging works, forming a complete closed loop from creation to exposure to feedback.

Comparison of AZ8 with Similar Competitors

Comparison Dimensions	AZ8	UpDream	LibTV
Product Positioning	Agentic Studio, driven by the Zumi agent on an infinite canvas, focused on AI video long-form narrative and multimodal creation.	A professional-grade AI video creation platform launched by Bilibili, designed around a complete creation workflow, emphasizing conversational creation and Skill accumulation.	A dual-entry AI video creation platform launched by LiblibAI, targeting both human creators and AI Agents, covering the full process from script to finished video.
Interaction Form	With the infinite canvas as the core, the Zumi Agent actively reads the canvas state and provides real-time suggestions for the next creation step.	Infinite canvas + node-based workbench, where creators describe ideas through dialogue and AI progressively breaks them down into script, storyboard, and finished video.	Infinite canvas + node-based workflow, where humans finely control each node through the GUI and Agents automatically orchestrate the process through the Skill interface.
Agent Capabilities	Built-in Zumi agent that analyzes project context in real time and proactively recommends creative actions, assisting long-video decision-making.	Long-memory AI assistant + multi-Agent collaboration, where the main Agent automatically splits complex tasks to sub-Agents for parallel processing.	Dual-entry design, where Agents can call platform capabilities through the open-source libtv-skills package, generating a 5-minute short drama in one sentence.
Model Strategy	Aggregates multiple external models such as Seedance, Wan, and MiniMax, allowing creators to switch flexibly as needed.	Integrates full official versions such as Seedance, emphasizing that each generation segment's information can be traced back to the official source.	Integrates 30+ mainstream models (Kling 3.0, Seedance 2.0, HappyHorse 1.0, etc.), with free switching within nodes.
Professional Control	3D clay models enable precise constraints on character actions and spatial relationships, suitable for complex narrative scenarios.	Skills on the canvas enable visualized workflows, where a single node can be modified independently and contextual style is automatically unified.	Provides 20+ professional film-grade features such as character three-view sheets, multi-camera 9-grid, lighting control, and 3D director's stage.
Editing Capabilities	Basic arrangement and asset management within the canvas, focusing on seamless connection in the generation stage.	Built-in editing pipeline in the canvas, allowing direct timeline arrangement, audio pairing, and exporting project files to DaVinci for fine-tuning.	Supports video composition and finished video delivery; the team edition provides a shared canvas and unified asset library for real-time multi-person collaboration.
Community Ecosystem	AZ8 Theater focuses on showcasing AI-native short films, building a space for creator exchange.	The Skill community accumulates a massive number of professional workflows, supporting creators in sharing and reusing others' methodologies.	Skill Hub provides 100+ professional Skills, while also relying on the Liblib community's 25 million creators to form a model ecosystem.

AZ8's Application Scenarios

AI-Native Short Film Creation: Creators use the infinite canvas and multi-model aggregation to conceive and produce complete AI short films from scratch.
Multi-Episode/Long Video Production: Orchestrate scripts, storyboards, characters, and scenes through Agent Zumi, reducing the difficulty of managing cross-shot consistency in long videos.
Character Consistency Animation: With the help of 3D precise control and a reference image system, maintain the continuity of the same character's appearance and actions across multiple scenes.
Advertising and Marketing Videos: Rapidly iterate commercial content such as brand promotional films and product demonstrations, with support for comparing multiple models to select the best one.
Creator Community Showcase: Publish works and get feedback through AZ8 Theater, forming a complete content ecosystem from creation to exposure.
Interactive Short Dramas/Narrative Content: Combine Agent-assisted script arrangement and storyboard generation to explore AI-driven forms of interactive storytelling.