Visa Crypto Labs has introduced a command-line payment tool that allows AI agents to execute transactions directly from terminal environments. The release points to growing demand for machine-to-machine payment infrastructure as automation expands across digital services.
The experimental product, called Visa CLI, enables programmatic card payments without requiring API key management. Users can authorize bots, scripts, and automated workflows to pay for services without manual input. The tool is currently in beta and requires a GitHub login for access, according to Visa’s documentation.
Excited to share Visa CLI, the first experimental product from Visa Crypto Labs. Check it out and request access here https://t.co/uxKKm2tB7y pic.twitter.com/7dh4Qckbtz
— Cuy Sheffield (@cuysheffield) March 18, 2026
Will AI Agents Drive The Next Payments Infrastructure Shift?
The launch reflects a broader push to enable autonomous financial interactions online. AI systems are increasingly handling tasks such as API consumption, cloud resource allocation, and service coordination. Traditional payment rails, built for human interaction, are not optimized for these use cases. By comparison, emerging protocols like x402 from Coinbase and Cloudflare, and the Machine Payments Protocol from Stripe and Tempo, are also targeting standardized frameworks for AI-native transactions.
But, Visa is positioning itself within this evolving stack rather than building in isolation. The company confirmed integration with Tempo’s Machine Payments Protocol, enabling card-based transactions within agent-driven workflows. This approach allows Visa to extend its existing network into new use cases without replacing core infrastructure.
“AI agents aren’t just answering questions. They’re executing tasks, coordinating services, consuming APIs, purchasing compute and completing automated B2B workflows,” Visa said in a blog post, adding that payment systems must adapt to support “command line commerce.”
Still, adoption will depend on how developers integrate these tools into real-world systems. Can command-line payments become a standard layer for automated services, or will competing protocols fragment the ecosystem? The next phase will hinge on whether large platforms and developers converge on shared standards for machine-driven transactions.
