PII Eraser

PII Eraser is a self-hosted API for detecting and redacting sensitive data in text and LLM chats. It is presented as a tool for PII anonymization and redaction, with examples showing names, locations, email addresses, phone numbers, and account numbers being replaced with safe labels.

Its features include OpenAI chat completions format support for creating guardrails, detection and redaction across 6 languages and 15 countries, and a set of localized entity types. The page also says it can run on mixed-language inputs and large documents, handle ingestion artifacts, and process entire documents and chat histories with linear compute scaling and no accuracy degradation. It describes redaction operators for replacing sensitive data with entity types, masking characters with asterisks, and cryptographic hashing for analytics. The site also notes support for over 1M tokens of context and performance above 5,000 tokens per second on an 8 vCPU instance, with optimizations for the latest x86 CPUs.

PII Eraser says it is built for regulated sectors, including financial services, legal, and insurance. Deployment runs entirely within a user’s infrastructure, either in a VPC or on-premises, in a single hardened container, and the page states that no external connections are required. The product page also references configuration of entity types via YAML in its documentation.

Pricing is described as unlimited usage licensing with no per-token fees, and the contracts are described as simple with no lengthy procurement. The site also links to docs, API reference, support, GitHub, release notes, security and compliance information, and product pages for pricing and features. Made in Europe appears in the footer.