Inspiration
AI agents are becoming a new class of internet consumer. Unlike humans, they often need only a specific paragraph, chart, citation, API response, or dataset row to complete a task.
NanoRail was built to help publishers monetize this new demand by giving agents a way to discover, evaluate, purchase, and access paid content through machine-payment infrastructure.
A key part of that vision is Tempo’s Machine Payments Protocol (MPP). Just as HTTP standardized how information moves across the web, MPP provides a standardized way for machines to discover paid resources, authorize payments, receive receipts, and access content. NanoRail applies those primitives to publisher monetization.
What it does
NanoRail is an SDK that lets publishers expose paid content to both humans and AI agents.
With a few data-nanorail-* attributes, NanoRail automatically provides:
- Content manifests
- A Publisher Discovery Agent
- Paid content unlocks
- Policy-controlled spending
- Receipts
- An Agent Wallet
- Activity dashboards
- Tempo-powered settlement
Agents discover content first, decide what to buy based on relevance and policy, then unlock only what they need. Human readers can also unlock premium content through seamless metered access.
Under the hood, NanoRail uses Tempo’s MPP infrastructure to support machine-native payment flows, authorization, receipts, and settlement, allowing content access to become a programmable economic interaction rather than a simple paywall.
How we built it
NanoRail is a TypeScript monorepo consisting of:
- Browser SDK
- Hono gateway
- Publisher demo
- Agent Wallet
- Activity dashboard
- Policy engine
- Receipt generation
- Tempo testnet settlement
The SDK scans publisher pages, generates manifests, and identifies monetizable content. Unlock requests go through policy checks, voucher verification, receipt issuance, and settlement.
Tempo’s MPP architecture serves as the foundation for the payment lifecycle. NanoRail builds on those primitives to add publisher discovery, policy-controlled spending, content metering, and user-facing agent controls.
Challenges
Our biggest challenge was balancing ambition with honesty.
We wanted to demonstrate a complete end-to-end flow—publisher integration, agent discovery, purchases, receipts, and settlement—without overstating the project's maturity.
Another challenge was designing a system that works for both humans and autonomous agents while keeping the developer experience simple.
We also had to bridge two worlds: publisher content systems and machine-payment infrastructure. Making Tempo’s MPP concepts understandable and useful in a publisher workflow required careful product and UX design.
Accomplishments
We're proud that NanoRail demonstrates a complete machine-consumption lifecycle:
- Working browser SDK
- Publisher Discovery Agent
- Policy-controlled agent spending
- Human and agent consumption flows
- Receipt generation
- Real Tempo testnet settlement
Most importantly, we turned the idea of publisher monetization for AI agents into a working prototype built on top of Tempo’s machine-payment infrastructure.
What we learned
We learned that agent payments are about more than moving money.
Discovery, policy enforcement, receipts, and user control are just as important as payment itself. Agents need structured ways to find content, while users need confidence that spending stays within defined limits.
We also learned that Tempo’s MPP primitives become significantly more powerful when combined with discovery and policy. Payment alone is not enough, agents need a way to find resources before purchasing them, and users need controls over what their agents are allowed to buy.
What's next
NanoRail is currently a working SDK and testnet MVP.
Next steps include:
- Publisher-owned content registration
- Durable storage
- Publisher accounts and authentication
- Multi-tenant deployments
- Improved receipt verification
- Simpler gateway deployment
Long term, we want NanoRail to support machine monetization across articles, APIs, datasets, research content, courses, and other digital resources while continuing to build on Tempo’s machine-payment ecosystem.
Built With
- github
- hono
- html/css
- mppx
- node.js
- npm
- railway
- react
- tempo
- typescript
- vercel
- viem
- vite
Log in or sign up for Devpost to join the conversation.