Insights, tutorials, and updates from the Proxy team
The top Claude Code and OpenClaw skills for giving AI agents payment capabilities. Install virtual card provisioning, spending controls, and transaction tracking in minutes.
Step-by-step tutorial for giving Claude Code payment capabilities using Proxy virtual cards, MCP integration, and spend policies.
Tutorial for configuring Proxy virtual cards in Cursor so your AI agent can make purchases, subscribe to services, and buy domains autonomously.
AI agents need dedicated virtual cards with spend controls, not access to your personal card. Here is why, and how the infrastructure works.
A head-to-head comparison of virtual card APIs for AI agent payments. Covers Stripe Issuing, Privacy.com, and Proxy across integration, controls, and agent readiness.
A technical overview of how AI agents make payments today. Covers the full payment flow, virtual cards, ACH/wire, crypto/stablecoins, and the x402 protocol.
A complete security checklist for AI agent payment access: card issuance, credential storage, MCP scoping, monitoring, revocation, and audit trails.
Claude, GPT, Cursor, and AutoGPT: which agents can actually support autonomous payments in 2026, and what controls are still required.
An implementation-focused comparison of virtual card APIs for AI agent use cases: issuance flow, controls, integration model, and operational fit.
Why corporate cards are a poor default for autonomous AI workloads and how prepaid virtual cards improve exposure control, compliance, and per-task issuance.
Set hard spending limits on AI agents using per-task and per-session controls with practical API patterns for autonomous payment safety.
Integrate Claude with MCP payment tooling using secure intent-based flows, scoped credentials, and verification-ready transaction logs.
A practical comparison of dedicated AI agent credit cards versus shared primary cards, including blast radius, auditability, and controls.
A practical comparison of virtual cards and virtual accounts for AI agents, including controls, acceptance, reconciliation, and rollout patterns.
A sourced market-sizing view of AI agent payments in 2026, including spend baselines, commerce signals, and scenario projections.
A data-driven roundup of the most important 2025-2026 stats on AI agent adoption, spending, and agentic commerce growth.
A 5-minute, copy/paste checkout flow for OpenClaw agents: declare intent, get isolated credentials, complete purchase, and verify the transaction.
Step-by-step guide to enabling AI agent payments safely. Learn why you shouldn't share your personal card, and how virtual cards with spending policies solve the problem.
A threat model and permission-scoping guide for OpenClaw payment skills: what to lock down, what to log, and what to never expose to an LLM.
Where to put guardrails in an OpenClaw checkout flow: hard limits, merchant locks, approval gates, and velocity controls that stop runaway spend.
A pattern for OpenClaw agents to manage SaaS subscriptions safely: billing-window unlock, merchant matching, amount tolerance, and clean reconciliation.
What
A buyer’s framework for evaluating agent payment platforms: rails, credential model, policy enforcement, verification, and operational auditability.
AI agents can accelerate procurement, reconciliation, and customer ops—but only if spending is isolated, auditable, and policy-bound.
How to make agent spend operationally safe: store intent-linked evidence records, capture receipts, and generate human-readable explanations for every charge.
Agents will buy the wrong thing. Build a refunds-and-returns loop that captures evidence, escalates when needed, and avoids disputes whenever possible.
Three ways to lock down where an agent can spend: MCC/category boundaries, exact-merchant allowlists, and descriptor matching. Tradeoffs and recommended defaults.
A practical comparison of agent payment rails: cards for universal acceptance, x402 for HTTP-native payments, and API billing for platform-specific spend.
A clean comparison of three credential models for agent spend: dedicated virtual cards, network tokenization, and shared/delegated tokens.
Token delegation exposes your entire credit line to agent failures. Dedicated cards create hard limits that policy engines cannot bypass.
Step-by-step tutorial for adding payments to AI agents using Proxy's MCP server with Claude Desktop or Cursor. Code examples included.
How AI agents authorized for one merchant category can silently spend elsewhere - and why dedicated cards are the only real defense.
Delegating your personal card to an AI agent sounds convenient, but it exposes your full credit line. Here's why dedicated cards are safer.
Create a virtual card with an isolated $50 balance for your AI agent. One API call, contained blast radius, no larger credit line at risk.
A comprehensive guide to agentic payments infrastructure in 2026: wallet providers, identity layers, card issuers, and network standards.
AI agents need payment rails that work today. Here's why virtual cards beat stablecoins for agentic commerce in 2026, and when that might change.
Learn the attestation-before-access pattern for AI agent payments - where agents declare intent before receiving credentials for safer autonomous spending.
A detailed post-mortem of a $4,800 AI agent overspend incident caused by retry loops and failed policy controls. Lessons for ops and finance teams.
As AI agents gain spending autonomy, chargeback rules built for humans face hard questions. We explore the liability gaps no one has solved yet.