Foil
Foil provides safety and growth infrastructure for the AI agent economy through two products: Detection, which gives companies device intelligence to identify AI agents, spoofed devices, and malicious behavior, and Agent Browser, a stealth Chromium browser with kernel-level patches that lets trusted agents navigate the web without being blocked. The company also develops the Agent Admission Protocol, an open standard enabling financial institutions to admit agents with explicit customer consent and institution-defined controls.
- Cybersecurity
- Artificial Intelligence
- AI Agents
- Financial Technology
- Data & Analytics
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
As AI agents increasingly perform tasks on behalf of users, businesses struggle to distinguish legitimate automated activity from malicious bots and fraud. Abusive agents exploit tampered browsers and synthetic traffic to commit fraud, while legitimate agents get blocked by anti-bot defenses, creating a dual challenge for companies and their customers.
Solution
Foil provides a two-sided platform for the AI agent economy. Its Detection product gives companies device intelligence to identify AI agents, spoofed devices, and malicious behavior through fingerprinting, behavioral analysis, and session replay. Its Agent Browser product offers a stealth Chromium browser with kernel-level patches and proxies that enables trusted agents to navigate the web without being blocked by CAPTCHAs or fingerprinting systems. The company also develops the Agent Admission Protocol, an open standard that lets financial institutions admit agents with explicit customer consent and institution-defined controls.
Target Audience
Primary customers are financial institutions, e-commerce platforms, and other businesses that need to protect against AI-driven fraud, as well as AI agent developers who need unblockable browser infrastructure for legitimate automation.
Features
- Device intelligence that detects spoofed browsers, creates stable unique fingerprints, and analyzes user behavior including taps, form fills, and mouse movements
- AI agent classification with hundreds of agent types identified, plus risk scoring for bot-like or human-like traffic
- Privacy-preserving session replay viewer that plays back suspicious sessions with detailed risk highlights
- Stealth Chromium browser with kernel-level patches and proxy integration to bypass CAPTCHAs and fingerprinting
- Agent Admission Protocol enabling agent-led account opening, lending, and payments with explicit customer consent and per-transaction limits
- SDKs for web and mobile platforms using platform-specific signals for maximum detection accuracy