Today’s AI-powered tools are fragmented, centralized, and monetized through opaque subscription models that don’t reward contributors or enable modular reuse. Users are locked into platforms, and developers building AI agents or microservices lack standardized ways to expose, monetize, or compose their work.

PayPerUse solves this by introducing a decentralized, composable AI agent marketplace — where every agent is modular, pay-per-use, and monetized natively using on-chain infrastructure.

The platform enables users to access intelligent services like research assistants, poetry generators, image creators, and even complex flows like NFT minting — all powered by autonomous AI agents and orchestrated via a usage-based payment protocol.

How it works:

• Each agent is a microservice with a defined input/output schema and is exposed via a pay-gated x402 HTTP API. This allows agents to be monetized per use — with no need for centralized billing, API keys, or subscriptions.

• Users fund their on-platform smart wallet (created via Coinbase’s CDP Wallet SDK), which programmatically authorizes payments when an agent is called.

• Payments are automatically split via CDP Wallet — between the agent creator, infrastructure providers, and optionally, affiliate referrers. The result is a fully on-chain, auditable, and autonomous revenue system.

Example: NFT Minting Agent

• A user enters a description and collection details.

• The agent calls an image generation model (via Amazon Bedrock), uploads the result and metadata to IPFS via Pinata, and mints an NFT on the Base chain — all in one click.

• Each step is orchestrated by autonomous, dedicated agents, executing sequentially and interacting via on-chain logic and pay-per-use flows.

• This agent showcases how PayPerUse enables composable, monetized AI workflows.

Akash Chat Integration

• Akash Chat is integrated into the platform to assist users in choosing the right agents and refining inputs before paying — ensuring that each agent call provides maximum value, especially when it’s metered and priced per request.

• We’ve also deployed a standalone Akash Chat Agent that enables users to interact with multiple decentralized LLMs such as Meta LLaMA, DeepSeek, and Qwen, providing censorship-resistant, low-cost, high-performance AI options through permissionless infrastructure.

Pinata for Storage

• Pinata is the default decentralized storage layer across the platform. All generated assets — including images, metadata, and other files — are stored on IPFS using Pinata.

• The Image Generation Agent includes an optional inline “Upload to IPFS” feature directly from the chat interface, allowing users to instantly store their generated content with a single action.

Create Custom Agentic Workflows

• Users can also create their own custom agentic workflows by composing multiple agents or services into a sequence — and list them on the marketplace for public use, with full control over monetization, pricing, and revenue sharing.

• Developers can create API endpoints which agent creators can utilize to compose agentic workflows across services and models.

• Workflows can include conditional logic, branching, and chained executions powered by pay-per-use models.

By combining modular AI agents, decentralized compute and storage, and a crypto-native monetization system, PayPerUse transforms how intelligent software is built, accessed, and paid for — unlocking a future where AI services are open, composable, and economically aligned.