Point One Percent FAQs

Question 1

What is Point One Percent and how does it secure AI agent commerce?

Accepted Answer

Point One Percent is a runtime security layer for AI agent commerce. It isolates payment credentials from AI agent prompts and logs, preventing unauthorized transactions by ensuring agents never directly access sensitive financial data, even during hallucinations.

Question 2

Can I use Point One Percent with my existing AI agent setup?

Accepted Answer

Yes, Point One Percent is designed to be compatible with any MCP (Multi-Modal Command Protocol) client. It provides straightforward integration instructions for popular platforms like Claude Code, Cursor, Windsurf, VS Code (Copilot), OpenClaw, and NemoClaw.

Question 3

What are the available Guardrail Modes in Point One Percent?

Accepted Answer

Point One Percent offers configurable guardrail modes: a keyword-based mode for zero-cost blocking of suspicious transactions, and an LLM-powered semantic analysis mode for more robust, AI-driven detection of malicious intent in high-value scenarios.

Question 4

Does Point One Percent support virtual cards or my own credit card?

Accepted Answer

Point One Percent supports multiple providers: 'Bring Your Own Card' (BYOC) using an encrypted local vault with CDP injection, integration with Stripe Issuing or Lithic for real virtual cards, and a 'Mock' provider for development and testing.

Question 5

How does Point One Percent prevent AI agents from accessing payment credentials?

Accepted Answer

It employs context isolation, keeping card credentials out of the AI agent's prompt context and logs. Payment data is stored in an AES-256-GCM encrypted vault and injected directly into the browser DOM via Chrome DevTools Protocol (CDP), completely bypassing the AI agent.

Point One Percent

Point One Percent

主な機能

ユースケース

主な機能

ユースケース