Blindfold

Detect, redact, tokenize, and mask PII across Python, JavaScript, Go, Java, .NET, CLI, and MCP. 80+ entity types, 30+ countries, works offline with zero dependencies.

Why Blindfold?

• Works offline, zero dependencies — No API key needed for local detection. No network calls. No external packages.
• 80+ PII entity types across 30+ countries with checksum validation (Luhn, IBAN mod-97, Verhoeff, etc.)
• 85x faster than Presidio — 0.4s vs 34s on 3,000 samples (benchmark)
• Higher accuracy — F1 58.6% vs Presidio 38.8% on AI4Privacy multilingual benchmark
• 8 operations: detect, redact, tokenize, detokenize, mask, hash, encrypt, synthesize
• Compliance-ready — Built-in GDPR, HIPAA, PCI-DSS policies
• Optional NLP upgrade — Add API key to detect names, addresses, organizations (60+ additional entities)
• 7 packages: Python SDK, JS/TS SDK, Go SDK, Java SDK, .NET SDK, CLI, MCP Server

Quick Comparison

Common Use Cases

• Sanitize LLM prompts — Strip PII before sending to OpenAI, Anthropic, etc.
• PII-safe RAG pipelines — Redact before embedding, restore after retrieval
• Log scrubbing — Anonymize data in logs and data pipelines
• GDPR/HIPAA compliance — Built-in policies for AI applications
• Synthetic test data — Format-preserving fake data generation

Packages

Quick Start (no API key needed)

JavaScript / TypeScript

import { Blindfold } from '@blindfold/sdk'

const client = new Blindfold()

// Detect PII locally — no API key, no network call
const result = await client.detect("Email john@acme.com, SSN 123-45-6789")
for (const entity of result.detected_entities) {
  console.log(`${entity.type}: ${entity.text} (score: ${entity.score})`)
}
// Email Address: john@acme.com (score: 0.95)
// Social Security Number: 123-45-6789 (score: 1.0)

// Redact PII locally
const redacted = await client.redact("Email john@acme.com, SSN 123-45-6789")
console.log(redacted.text)
// "Email, SSN"

Python

from blindfold import Blindfold

client = Blindfold()

# Detect PII locally — no API key, no network call
result = client.detect("Email john@acme.com, SSN 123-45-6789")
for entity in result.detected_entities:
    print(f"{entity.type}: {entity.text} (score: {entity.score})")
# Email Address: john@acme.com (score: 0.95)
# Social Security Number: 123-45-6789 (score: 1.0)

# Redact PII locally
result = client.redact("Email john@acme.com, SSN 123-45-6789")
print(result.text)
# "Email, SSN"

Java

import dev.blindfold.sdk.Blindfold;
import dev.blindfold.sdk.models.*;

Blindfold client = new Blindfold();

// Detect PII locally — no API key, no network call
DetectResponse result = client.detect("Email john@acme.com, SSN 123-45-6789");
for (DetectedEntity entity : result.getDetectedEntities()) {
    System.out.println(entity.getType() + ": " + entity.getText()
        + " (score: " + entity.getScore() + ")");
}

// Redact PII locally
RedactResponse redacted = client.redact("Email john@acme.com, SSN 123-45-6789");
System.out.println(redacted.getText());

Go

package main

import (
    "context"
    "fmt"

    blindfold "github.com/blindfold-dev/Blindfold/packages/go-sdk"
)

func main() {
    client := blindfold.New()
    ctx := context.Background()

    // Detect PII locally — no API key, no network call
    result, _ := client.Detect(ctx, "Email john@acme.com, SSN 123-45-6789")
    for _, entity := range result.DetectedEntities {
        fmt.Printf("%s: %s (score: %.2f)\n", entity.Type, entity.Text, entity.Score)
    }
    // Email Address: john@acme.com (score: 0.95)
    // Social Security Number: 123-45-6789 (score: 1.00)

    // Redact PII locally
    redacted, _ := client.Redact(ctx, "Email john@acme.com, SSN 123-45-6789")
    fmt.Println(redacted.Text)
    // "Email, SSN"
}

C# / .NET

using Blindfold.Sdk;

using var client = new BlindfoldClient();

// Detect PII locally — no API key, no network call
var result = await client.DetectAsync("Email john@acme.com, SSN 123-45-6789");
foreach (var entity in result.DetectedEntities)
{
    Console.WriteLine($"{entity.Type}: {entity.Text} (score: {entity.Score:F2})");
}
// Email Address: john@acme.com (score: 0.95)
// Social Security Number: 123-45-6789 (score: 1.00)

// Redact PII locally
var redacted = await client.RedactAsync("Email john@acme.com, SSN 123-45-6789");
Console.WriteLine(redacted.Text);
// "Email, SSN"

CLI

# No install needed
npx -y @blindfold/cli detect "Email john@acme.com, SSN 123-45-6789"

# Or install globally
npm install -g @blindfold/cli
blindfold redact "Email john@acme.com, SSN 123-45-6789"
blindfold redact --file sensitive-data.txt --quiet > clean.txt

MCP Server (Claude, Cursor, etc.)

Add to your claude_desktop_config.json, .mcp.json, or .cursor/mcp.json:

{
  "mcpServers": {
    "blindfold": {
      "command": "npx",
      "args": ["-y", "@blindfold/mcp-server"]
    }
  }
}

Then ask Claude: "Tokenize this patient record before summarizing: John Doe, SSN 123-45-6789"

Upgrade to Blindfold API (optional)

For names, addresses, organizations, and 60+ entity types, add your API key:

1. Sign up at blindfold.dev
2. Get your API key at app.blindfold.dev/api-keys
3. Set environment variable: BLINDFOLD_API_KEY=sk-***

// JS — auto-switches to NLP-powered API
const client = new Blindfold({ apiKey: 'sk-...', region: 'eu' })
const response = await client.tokenize("Contact John Doe at john@example.com")
// "Contact <Person_1> at <Email Address_1>"

# Python — auto-switches to NLP-powered API
client = Blindfold(api_key="sk-...", region="eu")
response = client.tokenize("Contact John Doe at john@example.com")
# "Contact <Person_1> at <Email Address_1>"

Operations

All operations work offline in local mode. All support batch processing (up to 100 texts per request).

Detection Policies

Regional Endpoints

See docs.blindfold.dev/essentials/regions for details.

Add your API key to unlock names, addresses, organizations, and 60+ additional entity types with NLP-powered detection.

Documentation

• docs.blindfold.dev — Full API reference and guides
• blindfold.dev — Product website
• status.blindfold.dev — Service status

License

MIT