Documentation — API, MCP Server, Office Add-in Setup

Q: What detection engines are used?

All packages include two detection engines: NLP Engine (Microsoft Presidio with spaCy for 25 languages, Stanza for 7 languages, XLM-RoBERTa for 16 languages) and Pattern Engine (regex-based with checksum validation). Use Hybrid Mode for maximum coverage.

Getting Started

Choose Your Path

We offer three packages: SaaS for quick start with pay-per-use, Managed Private for dedicated infrastructure, and Self-Managed for maximum control. All packages include 320+ entity types, 48 languages, and both NLP + Pattern detection engines. Request a quote to get started.

SaaS Package

Hosted on EU infrastructure with pay-per-use pricing. Includes API, MCP Server, Office Add-in, Desktop App, and Chrome Extension. 320+ entity types with 48 detection languages.

Self-Managed Package

Zero-knowledge architecture with full source code access. Air-gapped option available. Docker deployment, perpetual license, 48 detection languages, NLP + Pattern engines included.

API Reference

API Documentation

The anonymize.solutions RESTful API uses JWT authentication and provides endpoints for text analysis, anonymization, deanonymization, and batch processing. API documentation is provided to customers upon onboarding.

Authentication

All API requests require a valid JWT token. Tokens are obtained via the authentication endpoint using your API credentials. Tokens include role-based claims for access control.

Show Authentication Example

# Step 1: Obtain JWT token
curl -X POST https://api.anonymize.solutions/auth/token \
  -H "Content-Type: application/json" \
  -d '{"apiKey": "your_api_key", "apiSecret": "your_secret"}'

# Response: {"token": "eyJhbGc...", "expiresIn": 3600}

# Step 2: Use token in API requests
curl -X POST https://api.anonymize.solutions/analyze \
  -H "Authorization: Bearer eyJhbGc..." \
  -H "Content-Type: application/json" \
  -d '{"text": "Contact John at john@example.com"}'
                            

Core Endpoints

Four primary endpoints for PII processing:

/analyze – Detect PII entities in text
/anonymize – Apply protection operators to detected PII
/deanonymize – Reverse anonymization (encrypted entities only)
/batch – Process multiple items (up to 5,000)

Access API Documentation

Full API documentation including request/response schemas, error codes, rate limits, and integration examples is provided during onboarding. Contact us to request access.

AI Integration

MCP Server

The MCP (Model Context Protocol) Server enables AI tools to access anonymization capabilities directly. Available with the Enterprise tier, it provides 7 tools via Streamable HTTP transport.

Configuration

The MCP Server can be launched via npx and configured for use with compatible AI tools including Claude Desktop, Cursor, Windsurf, and other MCP-compatible clients.

                            // MCP client configuration example

                            {

                              "mcpServers": {

                                "anonymize": {

                                  "command": "npx",

                                  "args": ["@anonymize/mcp-server"]

                                }

                              }

                            }

Compatible Tools

The MCP Server integrates with any tool that supports the Model Context Protocol:

Claude Desktop
Cursor
Windsurf
Other MCP-compatible AI clients

Applications

Desktop App, Office Add-in & Chrome Extension

Desktop App

Built with Tauri 2.x for Windows, macOS, and Linux. Provides offline-capable anonymization with local vault encryption for storing anonymization mappings securely on the user's device.

Office Add-in

Anonymize PII directly within Microsoft Word and Excel. The add-in integrates into the Office ribbon and connects to your anonymize.solutions instance for real-time PII detection and protection.

Chrome Extension

Protect PII when interacting with AI chatbots and web-based tools. The extension detects and anonymizes sensitive data before it is sent to external services, with automatic deanonymization of responses.

FAQ

Frequently Asked Questions

How do I get access to the platform?

This is a managed hosting service. To get started, request a quote with your requirements. Our team will assess your needs and provision a dedicated environment for your organization.

Is my text data stored on your servers?

No. All text submitted for anonymization is processed exclusively in-memory and is never stored persistently. Input and output text data is discarded immediately upon completion of processing.

What detection engines are used?

All packages include two detection engines: NLP Engine (Microsoft Presidio with spaCy for 25 languages, Stanza for 7 languages, XLM-RoBERTa for 16 languages) and Pattern Engine (regex-based with checksum validation). Use Hybrid Mode for maximum coverage.

Can I deploy on my own servers?

The Enterprise tier supports on-premise deployment in addition to managed hosting. This includes full infrastructure documentation for Linux-based deployment with failover, load balancing, VPN, and PostgreSQL replication.

What protection operators are available?

All packages provide 7 protection operators: Replace, Redact, Mask, Hash, Encrypt, Asymmetric Encrypt (RSA-4096), and Keep. These can be configured per entity type and combined with compliance presets for regulatory requirements (e.g., GDPR).

Where is the infrastructure hosted?

All managed hosting infrastructure is located in Germany with Hetzner Online GmbH (ISO 27001 certified). No data is transferred outside the European Union.

Need Help Getting Started?

Request Demo Compare Packages