Poie is an industrial-grade AI asset fabrication platform. Think of it as a foundry where you can specify, cast, and deploy AI-generated assets using 50+ models in a structured workflow environment—no code required.

Units are our metered capacity for AI fabrication. Different models consume different amounts of units based on their complexity and throughput. For example, a fast image model costs 6 units, while a high-fidelity video model might require 934 units. Apprentice tier gets 500 units/month, and paid tiers get more with rollover.

We offer 50+ AI models across text, image, and video categories. This includes GPT-4o, Claude, Gemini for text; Flux, SDXL, Ideogram for images; and Kling, Runway, Luma for video. All models are accessible through the same unified interface. Check out our Models page for the complete list.

Yes! Poie supports multi-operator sync. You can see your teammates' activity, share blueprints, and work together on the same site simultaneously. This is ideal for production teams, agencies, and anyone who needs coordinated asset fabrication.

Yes, Pro and Agency plans include API access. You can run your saved workflows programmatically, integrate with your existing tools, and automate your AI pipelines. Enterprise customers can also get custom API solutions.

You can always upgrade your tier or purchase additional unit packs. On paid tiers, unused units roll over to the next month, so you won't lose what you've paid for. We'll also notify you when you're running low so you're never caught off guard.