Poolday is an applied research system for autonomous video editing. We are developing Creator-1, a novel video infrastructure platform. Creator-1 is a state-of-the-art video editing agent capable of interpreting users' natural language prompts to:
- understand your assets (image, video, audio, text)
- call any third party genAI models
- execute complex video edits end to end to assemble a complete video composition.
"After extensive evaluation, Poolday.ai stands out as the most intelligent and flexible video editing agent we've seen. Hands down the best in the market."
Our multi-agent system.
The platform currently operates with tens of specialized agents, each responsible for a specific capability such as editing logic, creative decisions, asset generation, audio, timing, structure, and quality control. New agents are added continuously.
Together, they function as a single system that plans, executes, and refines edits with minimal human input.
Agent: UI or API
Poolday's agents are available in two forms:
Poolday Interface
Agents operate within a full video editing environment and produce editable projects.
API Integration
Agents can be triggered from external systems such as Slack, internal dashboards, or task managers. You can still continue to edit the project manually after the agents are done.
Both methods use the same underlying architecture and generate fully editable outputs.
The output is not just a static video file, but a fully editable project.
Agents cut, trim, translate, generate AI assets, replace scenes, apply styles, and orchestrate over 100 generative models within one workflow. The result remains fully editable, allowing human intervention only when needed.
The system is designed to be extended by its users.
Users can load custom skills into the platform with simple text inputs, no code required, including logic, constraints, or creative rules, so the agent network edits according to their specific needs. The architecture is built to adapt, not to enforce a fixed workflow.
Progress is measured by autonomy.
Our primary metric is the ratio of editing actions executed by the agent versus the human for a given video. Each release aims to reduce manual interaction and increase the portion of a video completed autonomously, moving steadily toward full agent-driven editing. For hundreds of identified use cases, our Autonomy Ratio is 100%.
Autonomy Ratio α = Aagent / Atotal
Current system average for straightforward video editing: 100% autonomous actions
Current system average across all video complexity levels: 64% autonomous actions
Request Access
Poolday is available by request to individuals and teams looking to use the system in experimental or production environments.
Early users influence agent design and gain access to new capabilities as they are introduced.
If you would like to participate, request access. Responses are sent within 12 hours, and access is expanded gradually as the system matures.