The payment + trust layer for MCP
MCP is quickly becoming the universal interface for AI tools. But paid tools introduce a hard problem: most clients can call tools, yet they can't reliably handle on-chain micro-payments, budgets, and audit trails.
via402 solves that with a client-agnostic payment and safety layer built around X402. You approve paid tools once, define rules (auto-pay, ask, or block), and get a provisioned MCP endpoint your agents can use from any MCP-compatible client—without forcing every user to manage a crypto wallet.
Project started
Started research on machine-friendly payments for tools and APIs, with a focus on standardized flows and strong auditability.
X402 + MCP direction
Aligned on HTTP 402 payment flows for tool calls and began designing a proxy-based approach that stays client-agnostic.
Avalanche Fuji beta rails
Shipped the first end-to-end payment rails on Avalanche Fuji using USDC-style flows, enabling real payment logic in a safe test environment.
Dashboard + via402 beta
Launched the via402 web app: accounts, a policy dashboard, transaction logs, and automated provisioning of user-specific MCP endpoints with X402 payment support.
Built for the next wave of on-chain applications
As agents become the UI, tool access turns into an economy. via402 makes that economy usable today—across clients—with policy controls, transparent accounting, and a security layer that helps teams trust what agents are calling.
Network supported
Avalanche Fuji today. More networks (including Base and Solana) are planned next.
1 (Beta)
Auto-pay rules
Approve, limit, or block spending per tool—so agents don't surprise you.
Per tool
Every payment tracked
See tool, amount, chain, timestamp, and status—everything in one dashboard.
Full audit log