Redact PII before it ever leaves your network.
anonymix.dev is a high-performance, self-hosted service that redacts PII (Personally Identifiable Information) and sensitive data in real time. Call it as a REST API, or drop it in as a redacting proxy that strips PII from requests before forwarding them upstream — no application code changes. Either way, confidential customer records never leave your network unredacted.
POST /v1/anonymize
Content-Type: application/json
{
"email": "john@example.com",
"ssn": "123-45-6789",
"text": "Contact john@example.com or call 555-123-4567. SSN on file: 123-45-6789."
}
→ 200 OK
{
"email": "[REDACTED]",
"ssn": "[REDACTED]",
"text": "Contact [REDACTED] or call [REDACTED]. SSN on file: [REDACTED]."
}Complete Control Over Sensitive Data Sanitization
Modern enterprise software architectures handle millions of data transactions every second across cloud services, internal microservices, and third-party vendors. Without an automated security layer, raw customer information can easily leak into log aggregators, analytical databases, or AI models. Built on a compiled Go backend, anonymix allows developers to automatically redact PII directly at the ingestion point, guaranteeing that sensitive payload data never leaves your network boundaries without proper anonymization.
Built for compliance-critical pipelines
Drop it into API gateways, ETL jobs, logging aggregators, and LLM pre-processing flows.
Ultra-Low Latency Engine
In-memory JSON and XML transforms backed by a highly optimized, compiled RE2 regex pipeline. Experience millisecond processing speeds with no cold starts and zero external API round-trips when you redact PII inside your local network.
Stateless Privacy Architecture
Our processing pipeline is 100% stateless. Absolutely nothing is ever written to local disk, persistent storage, cache, or third-party cloud databases. Sensitive information never leaves your network perimeter during the request lifecycle.
Self-Hosted & On-Premises Deployment
Deploy completely inside your own security perimeter. Run anonymix as a bare-metal binary, inside your virtual private cloud (VPC), air-gapped enterprise environments, or local Kubernetes clusters.
Deterministic & ML Accuracy
Leverage fast, deterministic path rules and regex (Tier 1) paired with an optional, high-performance pure-Go Named Entity Recognition (NER) pipeline (Tier 2) to accurately redact PII from unstructured text and complex payloads.
Fully Customizable Masking Rules
Configure exactly how your application handles sensitive data. Customize replacement tokens, masking rules, custom regex patterns, and specific JSON/XML fields dynamically using YAML configurations and environment variables.
Two Modes: REST API or Drop-In Proxy
Run anonymix as a clean HTTP REST API you call from any language, or as a redacting proxy: point an existing client at anonymix, set one header, and PII is stripped from the request body before it reaches the upstream — zero application code changes. Achieve immediate GDPR, CCPA, and HIPAA compliance with absolutely zero external package dependencies.
Secure your data flows in 3 simple steps
Cleanse incoming application logs, webhook payloads, and analytics requests of sensitive customer identifiers before sending data downstream to databases, cloud storage, or AI providers.
01
Deploy On-Premises
Spin up the lightweight anonymix binary inside your private cloud VPC, local Kubernetes cluster, or air-gapped infrastructure right alongside your backend microservices.
02
Route Your Payloads
Either POST raw JSON or XML to the local REST endpoint, or run anonymix in proxy mode and point an existing client at it with the destination named in one header. Detection runs entirely on your own hardware.
03
Clean Downstream Writing
In REST mode you get a redacted, schema-intact payload back in milliseconds. In proxy mode anonymix forwards it for you, so third-party APIs, LLM prompts, analytics platforms, and log aggregators only ever see redacted data.
Comprehensive PII & Sensitive Entity Detection
anonymix scans both structured fields and unstructured text blocks to automatically sanitize a wide variety of personal data entities.
Direct Identifiers
Full Names, Social Security Numbers (SSN), Passport Numbers, Driver's License IDs, Tax Identifiers, and National Identity Numbers.
Contact & Location Data
Email Addresses, Global Phone Numbers, Physical Street Addresses, ZIP/Postal Codes, GPS Coordinates, and Geographic Subdivisions.
Network & Device Identifiers
IPv4 & IPv6 Addresses, MAC Addresses, User-Agent Strings, Device UUIDs, Session Tokens, and Client Connection Logs.
Financial Data & Credentials
Credit & Debit Card Numbers (Luhn Validated), IBAN Numbers, Bank Account Details, API Secret Keys, Passwords, and Authentication Tokens.
Enterprise Use Cases for the Redaction Proxy
Explore how engineering teams utilize local PII masking to protect their data infrastructure.
LLM & AI Pipeline Redaction
Route prompt payloads and document chunks bound for third-party AI APIs like OpenAI or Anthropic through anonymix in proxy mode. Sensitive fields are redacted in flight, so confidential customer data never leaves your network — and the calling service keeps its existing request code.
Centralized Log Aggregation
Put anonymix in front of your log shipper's egress path to cleanse application logs before they reach external platforms like Datadog, Splunk, or Elastic. Prevent developers from accidentally leaking plain-text customer credentials, SSNs, or payment details.
Outbound Webhook & Third-Party Integration Filtering
Forward outgoing webhook events through anonymix to strip PII before they hit external endpoints. Maintain strict regulatory boundaries without touching the sender's serialization or your underlying database architecture.
Frequently Asked Questions
Out of the box, anonymix identifies standard global identifiers such as email addresses, IP addresses, credit card numbers, Social Security Numbers (SSNs), phone numbers, and physical addresses. You can easily extend this to redact PII specific to your domain by defining custom regex patterns or targeting specific JSON keys and XML tags via YAML configuration files.
Yes — that is what proxy mode is for. Start anonymix with mode: "proxy", whitelist the upstream origins it may forward to, then point your client at the anonymix instance and set the X-Anonymix-Target header to the real destination URL. anonymix redacts the request body and forwards the request on with the method, headers, and schema intact, relaying the upstream's response back unmodified. The only change on the caller's side is the base URL and one header.
It works with any HTTP upstream that accepts JSON or XML, including LLM provider APIs, webhooks, CRMs, and analytics endpoints. Note that anonymix is not an OpenAI-compatible drop-in: it does not emulate the /v1/chat/completions surface, so you point your client at anonymix and name the real destination in the X-Anonymix-Target header rather than only swapping a base URL. Destinations must be explicitly whitelisted, and connections to loopback, private, and link-local addresses are blocked at dial time to prevent SSRF.
Because anonymix runs entirely within your private network or cloud VPC, sensitive customer data never leaves your network perimeter to reach a third-party SaaS processor. This drastically reduces your compliance audit footprint, keeping data pipelines thoroughly secure and compliant under strict regulatory frameworks.
We designed anonymix in Go specifically for high-throughput, low-latency workloads. Because the API runs locally alongside your database, backend microservices, or API gateway, you eliminate the high network round-trip latencies associated with public cloud PII APIs, keeping processing overhead under a single millisecond.
Yes! The core anonymix binary has zero runtime external dependencies and does not make calls to any external servers. It is built from the ground up to operate seamlessly in completely isolated, air-gapped government, healthcare, or enterprise networks.
anonymix performs recursive AST parsing on JSON objects and XML nodes. It navigates deeply nested arrays, objects, and attributes to identify and redact PII while preserving the original schema and structural integrity of your payloads.
Ready to redact sensitive data on-premises?
Paste a JSON or XML payload into our live interactive playground and experience real-time PII sanitization in action before data leaves your network.
Open the live demo