API Playground — Test Watermark Removal Live

Experiment with the Claude Watermark Remover API directly from your browser. Select an endpoint, configure your request, and see the response in real time with syntax highlighting and performance metrics. No API key required — the playground calls the live, open API directly from your browser.

Request

Content-Type: application/json

Response 

Click "Send Request" to see the response here.

The playground simulates real API responses with
realistic data, timing, and response structure.

Try different endpoints and request bodies to
explore the full API surface.

Example Requests to Try

The playground comes preloaded with sample text that contains simulated watermark patterns. You can modify the request body to test with your own content, switch between endpoints, and experiment with different model settings. Below are some scenarios worth exploring to understand the API's capabilities and response format.

Text Removal — Standard

The default mode. Balances watermark removal effectiveness with minimal text modification. Ideal for blog posts, articles, and general content where you want clean output without aggressive rewriting. Typically modifies 15-25 tokens per 1000 words of watermarked content and achieves removal confidence above 90%.

json
{
  "syn-string">"text": "syn-string">"Your watermarked text...",
  "syn-string">"model": "syn-string">"standard"
}

Text Removal — Aggressive

Maximum strength removal. Substitutes more tokens and applies broader distribution normalization. Best for content that will be analyzed by watermark detectors, where you need to ensure zero false positives. May modify 30-50 tokens per 1000 words but maintains semantic equivalence throughout.

json
{
  "syn-string">"text": "syn-string">"Your watermarked text...",
  "syn-string">"model": "syn-string">"aggressive",
  "syn-string">"include_diff": true
}

Detection Only

Analyze text for watermark presence without removing anything. Returns a confidence score and statistical test results. Useful for building content classification systems, moderation queues, or audit trails where you need to flag AI-generated content without altering it.

json
{
  "syn-string">"text": "syn-string">"Text to analyze for watermarks..."
}

With Diff Output

Enable the diff feature to see exactly which tokens were modified and why. The response includes a detailed array showing original tokens, their replacements, and the reason each substitution was made. Essential for transparency, debugging, and quality assurance workflows.

json
{
  "syn-string">"text": "syn-string">"Your watermarked text...",
  "syn-string">"include_diff": true,
  "syn-string">"model": "syn-string">"conservative"
}