Agent Commerce Infrastructure Live
By AREMES & Ryan Seslow
March 22nd 2026
This system enables autonomous AI agents to discover, evaluate, and license artwork directly from a living archive. The first experiment, my own archive!
It is built on existing infrastructure across ryanseslow.com and aremes-enterprises.com. No third-party marketplace is required and no manual intervention is needed. This is a machine-readable and machine-transactable creative system. What exists right now is a fully operational foundation. The archive includes over 1,000 blog posts indexed as structured VisualArtwork entries and more than 9,000 media library images that are accessible and queryable. A public catalog.json endpoint exposes the archive to external systems, supported by a REST API with pagination and filtering so agents can navigate the work efficiently.
Across both sites, JSON-LD schema is embedded into every page and product. This includes CreativeWork, Product, Organization, and Service schema types, making the entire system legible to AI agents and machine crawlers. AREMES functions as the agent interface layer. It provides an agent.json endpoint that describes available services, licensing structures, and commission pathways. It also connects both domains into a unified system so that discovery on one leads naturally to the other.
The discovery layer is active through .well-known endpoints, openapi.json, and llms.txt files. These allow AI agents and language models to identify, interpret, and interact with the system without human assistance. Two payment rails are fully operational. Stripe supports purchase-intent requests that generate live checkout URLs, create orders programmatically, and confirm transactions through webhooks. In parallel, the x402 protocol enables native machine-to-machine payments using USDC on the Base network. An agent can request a quote, send payment on-chain, and verify the transaction using a hash, with funds settling directly to the wallet at minimal cost. All transactions are verified through a blockchain layer using Alchemy RPC, ensuring that payments are confirmed automatically and reliably. Orders are tracked in WooCommerce and tagged with metadata identifying them as agent-initiated, including agent IDs where applicable.
An AI agent can complete a full transaction cycle without any human involvement. It can discover the system through .well-known endpoints, read the archive via catalog.json, filter and select a work, send a purchase request, receive payment instructions, complete the transaction, and receive confirmation and access to the licensed asset. This process has already been executed successfully within the system. An agent initiated a purchase request, generated a WooCommerce order, and completed the transaction flow with proper tagging and confirmation. This validates that the infrastructure is functional end to end.
The scale of the system is significant. It represents over 50 years of creative output, with thousands of works already accessible and many more still being added. The archive is not static. It grows continuously as new work is published. This enables a new model where artwork is not only viewed or purchased by humans but discovered and licensed by machines. It allows for training data access, programmatic purchasing, and continuous machine-driven discovery.
In parallel with building this system for my own work, I am now offering this exact infrastructure as a service for other artists, designers, agencies, and creative professionals. Most people have valuable archives that are invisible to AI systems. I build the bridge that makes their work machine-readable, discoverable, and transactable in the emerging agent economy.
The infrastructure is live. The interface is open. The next phase is external activation as agents begin to discover, interact with, and transact through the system independently. This is a new model of creative distribution and value exchange. A human archive, structured for machine understanding, capable of autonomous transaction, built in March 2026 on top of a 50-year body of work.
AREMES -> https://aremes-enterprises.com/
RSMAD -> https://www.ryanseslow.com/the-ryan-seslow-museum-of-art/
About the art works in this post – I used google gemini’s nano-banana 2.5 to render several of my existing illustrations and paintings into enamel pins. I may still turn some of these into actual pins, but it was more fun to bring those renderings into adobe photoshop to create various assets and recompose them into new illustrations. The 4 images in this post were the first of their kind on March of 2026.
