platoseedplatoseed
AI-powered fraud investigations
MouseCat enables companies to automatically detect, investigate, and mitigate emerging fraud trends
MouseCat is an AI-powered platform for fraud investigations designed to scale risk operations by deploying within a customer's own cloud environment. It aims to automate and improve human-quality investigations, enabling faster decisions and proactive fraud discovery.
Deploys entirely within customer cloud accounts, integrating with the team’s data, vendors, and APIs. Provides agents and an infrastructure stack for building, evaluating, and tuning autonomous investigation workflows, with case-level feedback and ground-truth labels to iteratively improve performance. Stores evidence, reasoning, and outputs for trend analysis and supports analysts building their own agents on top of MouseCat’s framework. Offers private beta results, including high precision and recall metrics for transaction reviews, and works across the customer’s data landscape to run investigations in queues or in the background.
Who it’s for: Risk teams and fraud investigations groups in enterprises that require scalable, in-cloud, customer-controlled fraud analysis and integration with existing data and vendor ecosystems.
Private beta results and traction highlighted (e.g., crypto exchange results, private deployments); founding team with YC backing and notable industry ties.
AmazonNicholas Aldridge is the Co-Founder and CEO of MouseCat, and a core maintainer of MCP. He spent 6.5 years as a Principal Engineer at AWS AI, where he helped launch and lead Amazon Bedrock Knowledge Bases, Agents, and AgentCore. He also represented Amazon on the A2A steering committee. Nicholas has helped some of the largest and most security and privacy sensitive organizations in the world to plan and implement their AI strategies.
Joseph McAllister is the Co-Founder and CTO of MouseCat. Before MouseCat, Joseph spent 4 years building systems that power every risk decision at Coinbase. He focused on streaming pipelines, large-scale data processing workloads, and improving ACH and ATO risk models. While studying Computer Science at Cornell University, he founded Roo Storage, which was acquired in 2020 by Handled (now UniGroup).
MouseCat does human-quality fraud investigation at scale
MouseCat provides an AI-powered platform for continuous, human-quality fraud investigation at scale. It integrates with customers’ data, learns from analyst feedback, and operates post-transaction to identify fraud patterns, propose mitigations, and continuously improve with on-prem data integration.
Formerly “Agnostic”
AI coworkers for fraud and risk teams.
Behavioral intelligence infrastructure for anti-fraud
