@aismarttalk/anondocs-sdk

AnonDocs SDK

TypeScript/JavaScript SDK for AnonDocs - Privacy-first text and document anonymization powered by LLMs.

Installation

npm install @aismarttalk/anondocs-sdk

Quick Start

import { AnonDocsClient } from '@aismarttalk/anondocs-sdk';

const client = new AnonDocsClient({
  baseUrl: 'http://localhost:3000',
  defaultProvider: 'ollama'
});

// Anonymize text
const result = await client.anonymizeText('My name is John Smith and my email is [e-mail]');

console.log(result.anonymizedText);
// Output: "My name is [NAME] and my email is [EMAIL]"

console.log(result.piiDetected);
// Output: { names: ['John Smith'], emails: ['[e-mail]'], ... }

Features

• 🔒 Privacy-first anonymization using local or cloud LLMs
• 📄 Document support (PDF, DOCX, TXT)
• 🌊 Streaming with real-time progress updates
• 🎯 Full TypeScript support with strict types
• 🚀 Works in Node.js and Browser
• ⚡ Multiple LLM providers (OpenAI, Anthropic, Ollama)

API Reference

Initialize Client

const client = new AnonDocsClient({
  baseUrl: 'http://localhost:3000',  // Optional, defaults to localhost
  defaultProvider: 'ollama',          // Optional: 'openai' | 'anthropic' | 'ollama'
  timeout: 30000                      // Optional, defaults to 30s
});

Check Health

const health = await client.health();
// Returns: { status: 'ok', timestamp: '2025-11-04T...' }

Anonymize Text

const result = await client.anonymizeText(
  'John Smith lives at 123 Main St and his email is [e-mail]',
  { provider: 'ollama' }  // Optional, uses defaultProvider if not specified
);

console.log(result.anonymizedText);
console.log(result.piiDetected);
console.log(result.processingTimeMs);
console.log(result.wordsPerMinute);
console.log(result.chunksProcessed);

Anonymize Document

Node.js:

import { readFileSync } from 'fs';

const fileBuffer = readFileSync('./document.pdf');
const result = await client.anonymizeDocument(fileBuffer, {
  provider: 'ollama'
});

Browser:

const fileInput = document.querySelector('input[type="file"]');
const file = fileInput.files[0];

const result = await client.anonymizeDocument(file, {
  provider: 'openai'
});

Streaming Anonymization

Get real-time progress updates for long texts or documents:

await client.streamAnonymizeText(longText, {
  provider: 'ollama',
  onProgress: (event) => {
    console.log(`${event.progress}% - ${event.message}`);
    // Progress: 0-100
    // Types: 'started', 'chunk_processing', 'chunk_completed', 'completed', 'error'
  },
  onComplete: (result) => {
    console.log('Done!', result.anonymizedText);
  },
  onError: (error) => {
    console.error('Error:', error.message);
  }
});

Streaming Document Anonymization

await client.streamAnonymizeDocument(file, {
  provider: 'ollama',
  onProgress: (event) => {
    updateProgressBar(event.progress);
  },
  onComplete: (result) => {
    displayResults(result);
  },
  onError: (error) => {
    showError(error);
  }
});

Response Types

AnonymizationResult

interface AnonymizationResult {
  anonymizedText: string;
  piiDetected: {
    names: string[];
    addresses: string[];
    emails: string[];
    phoneNumbers: string[];
    dates: string[];
    organizations: string[];
    other: string[];
  };
  chunksProcessed: number;
  wordsPerMinute: number;
  processingTimeMs: number;
}

ProgressEvent

interface ProgressEvent {
  type: 'started' | 'chunk_processing' | 'chunk_completed' | 'completed' | 'error';
  progress: number;        // 0-100
  message: string;
  currentChunk?: number;   // 1-indexed
  totalChunks?: number;
  data?: AnonymizationResult;  // Only on 'completed'
}

Error Handling

The SDK provides typed error classes:

import { 
  AnonDocsError,
  AnonDocsApiError,
  AnonDocsNetworkError,
  AnonDocsValidationError,
  AnonDocsStreamError
} from '@aismarttalk/anondocs-sdk';

try {
  await client.anonymizeText('');
} catch (error) {
  if (error instanceof AnonDocsValidationError) {
    console.error('Validation error:', error.message);
  } else if (error instanceof AnonDocsApiError) {
    console.error('API error:', error.statusCode, error.message);
  } else if (error instanceof AnonDocsNetworkError) {
    console.error('Network error:', error.message);
  }
}

Supported File Types

• PDF: application/pdf
• DOCX: application/vnd.openxmlformats-officedocument.wordprocessingml.document
• TXT: text/plain

Max file size: 10MB

LLM Providers

• OpenAI - Cloud-based, requires API key
• Anthropic - Cloud-based, requires API key
• Ollama - Local/self-hosted, privacy-first

Examples

See the examples/ directory for complete usage examples:

• basic-usage.ts - Simple text anonymization
• streaming-usage.ts - Real-time progress tracking
• document-usage.ts - File upload for Node.js and Browser

TypeScript Support

The SDK is written in TypeScript and includes full type definitions. All types are exported:

import type {
  LLMProvider,
  PIIDetected,
  AnonymizationResult,
  ProgressEvent,
  ClientConfig
} from '@aismarttalk/anondocs-sdk';

Requirements

• Node.js: >= 18.0.0 (uses native fetch, FormData, Blob)
• Browser: Modern browsers with fetch and ReadableStream support
• Zero runtime dependencies - works everywhere!

License

MIT

Links

• 🌐 Website: anondocs.aismarttalk.tech
• 📦 NPM: @aismarttalk/anondocs-sdk
• 💻 GitHub: AI-Smarttalk/anondocs-sdk
• 📚 Documentation: anondocs.aismarttalk.tech/documentation

Made with ❤️ by AI SmartTalk