PEyeEye AI is a platform designed to protect sensitive customer data when integrating large language models (LLMs) into applications. It addresses the challenge of preventing personally identifiable information (PII) from being exposed to third-party AI models by redacting PII from prompts before they are sent to an LLM and rehydrating the redacted data in responses, all within a single round-trip. The system uses deterministic tokens to replace sensitive information, ensuring the mapping remains consistent within a session while remaining random across sessions. Zero data retention is enforced by default, with stateless AEAD-sealed sessions and AES-256-GCM encryption at rest, so no sensitive data is stored during processing.
The platform detects 62 types of PII entities out of the box, including names, emails, phone numbers, credit cards, SSNs, addresses, bank accounts, medical record numbers, IP addresses, license plates, passport numbers, geo coordinates, company names, order IDs, currency amounts, API keys, and tax IDs. Users can also define custom detectors by providing positive and negative examples, allowing the tool to recognize domain-specific PII without the need for custom model training or data science intervention. Detection is performed using a hybrid approach that combines regular expressions and machine learning-based named entity recognition, supporting a wide range of use cases such as support tickets, medical notes, and financial transactions.
PEyeEye AI is delivered as an API with two primary endpoints for redaction and rehydration, and offers SDKs for multiple programming languages including Python, TypeScript, and Go. shield() call, or developers can interact directly with the API. The service supports streaming with server-sent events and partial-token buffering, making it suitable for real-time applications. The platform emphasizes composability, allowing users to customize detection policies and entity types as needed. 3 for data in transit and opinionated defaults for privacy protection.
Designed for developers and organizations deploying AI features that handle sensitive user data, PEyeEye AI provides a stateless, privacy-preserving solution to safely leverage LLM capabilities without leaking confidential information.
In the Other AI space, PEyeEye AI takes a focused approach. It focuses on preventing sensitive PII from leaking when sending prompts to and from large language models. PEyeEye AI is a B2B product aimed at developers integrating LLMs and handling sensitive data. There is a free tier. It runs on the command line and API.
PEyeEye AI first shipped in 2026. The project is developed in the open on GitHub with 9 commits in the last 90 days. Across PulseGate's embedding index, PEyeEye AI has few near neighbours, marking it as relatively distinct. Among its 6 catalogued features are PII redaction, PII rehydration, and API access. It exposes integrations via a public API.
Latest indexed changes and source events
peyeeye.ai discovered by the PulseGate indexer
Other apps tracked under the same category.
