Why FareSide Exists

We didn’t plan to start a company around x402 infrastructure. We built x402-rs, our Rust implementation of x402 protocol, because the protocol itself was interesting, Rust was the right tool to handle volume of projected billions of AI agents, and the ecosystem needed an open-source implementation that worked.

It did work better than expected.

The Surge

We started with a simple facilitator for developers to test against. Then Polygon reached out with a PR to support their network. That collaboration led us down the path of making it actually scalable: managing nonces across multiple private keys, handling concurrent transactions, building defensive architecture that wouldn’t fall apart under real load.

Then the memecoin frenzy hit.

x402-rs became the second-largest facilitator by transaction volume, briefly hitting 77% of all x402 traffic. The exploration paid off: our infrastructure stayed online while others went down. But it also proved exactly how big the operational can of worms actually is at scale.

Gas spikes during high activity. Nonces collide across instances. RPC providers rate-limit you when you need them most. Transactions get stuck and need manual intervention. Every edge case you didn’t plan for shows up in production, usually at 3am.

We handled it. Then we shut it down.

Not because it failed. Because subsidizing hundreds of thousands of daily transactions out of pocket isn’t a business model. We weren’t chasing clout or building for a leaderboard. We’d proven the infrastructure could handle real load. The economics couldn’t.

What We Learned

Running a facilitator at scale taught us things the spec doesn’t cover:

Settlement is probabilistic. Gas spikes. Nonces collide. Transactions get stuck. RPC providers rate-limit you at the worst possible time. You need defensive architecture and redundancy at every place.

Operations matter more than features. The difference between a facilitator that works in demos and one that works in production is whether transactions actually land on-chain when the network is congested and your private key management doesn’t cause nonce disasters across multiple instances.

Sustainable infrastructure needs sustainable economics. Free facilitators are marketing tools. They work until they don’t, and their priorities shift with whoever’s paying for them. Production workloads need infrastructure that’s aligned with your operational requirements, not someone else’s promo budget.

Infrastructure failures break revenue. Every failed settlement is a lost dollar and lost trust you never get to retry.

The Fragmentation Problem

While we were rebuilding, we have been watching the ecosystem split.

One group strives to build own walled gardens with their own clients, their own APIs, incompatible with the broader protocol. Slightly better UX, marginally valuable features, but ultimately N-squared integration complexity.

Another group doubled down on the protocol: x402 as the standard, interoperability as the goal, Coinbase and Cloudflare backing it with real resources. Slower to mature, but if it works, any compliant client talks to any compliant API.

It all did play out before with TCP/IP vs IPX. The story is being unfolded now with stablecoins eating the lunch of banksters. Open protocols compound, proprietary platforms plateau.

What We’re Building

FareSide is us bringing x402-rs back as production infrastructure, operated sustainably.

We rebuilt the transaction relayer with hard-won lessons about failure isolation, observability, and operational discipline. The same Rust codebase now powers multiple production deployments, including Polygon’s official facilitator, UltravioletaDAO, x402labs and countless other folks. We know it works because other people are running it at scale.

We’re not trying to own the market. We’re trying to provide one viable option for teams that need production x402 infrastructure but don’t want to run it themselves. Protocol-first, multi-chain, interoperable by design.

What’s Next

We’re in private beta, working with select partners to validate production requirements. We’re making sure the infrastructure works before we open it up.

If you’re building something real on x402: an API that agents will pay for, a service that needs reliable settlement, and x402 is on your critical path to revenue, we’d like to work with you. We are here to be the boring, reliable layer you never have to think about again.

Join the waitlist or self-host with x402-rs if you want to do it all on your own.