Dogwhistle Detection

Identify coded language that appears innocuous but carries hidden harmful meanings.

What it detects:

Coded language: Terms with hidden meanings to specific groups
Subtlety level: How hidden the messaging is
Harm potential: Risk level of the detected content
Categories: Types of dogwhistles detected

Reference

See the Dogwhistle Detection API Reference for method signatures and parameters.

Response: DogwhistleResult

By Example

Basic Usage

Python
PHP
REST API

Blocking Client
Async Client

result = client.check_dogwhistle(
    "This comment to analyze for coded language",
    article_id
)

print(f"Detected: {result.detection.dogwhistles_detected}")
print(f"Confidence: {result.detection.confidence:.2f}")
print(f"Reasoning: {result.detection.reasoning}")

if result.details:
    print(f"Terms: {result.details.dogwhistle_terms}")
    print(f"Categories: {result.details.categories}")
    print(f"Subtlety: {result.details.subtlety_level:.2f}")
    print(f"Harm potential: {result.details.harm_potential:.2f}")

result = await client.check_dogwhistle(
    "This comment to analyze for coded language",
    article_id
)

print(f"Detected: {result.detection.dogwhistles_detected}")
print(f"Confidence: {result.detection.confidence:.2f}")
print(f"Reasoning: {result.detection.reasoning}")

if result.details:
    print(f"Terms: {result.details.dogwhistle_terms}")
    print(f"Categories: {result.details.categories}")
    print(f"Subtlety: {result.details.subtlety_level:.2f}")
    print(f"Harm potential: {result.details.harm_potential:.2f}")

$client->checkDogwhistle($articleId, "This comment to analyze")
    ->then(function ($result) {
        echo "Detected: " . ($result->detection->dogwhistlesDetected ? 'Yes' : 'No') . "\n";
        echo "Confidence: " . number_format($result->detection->confidence, 2) . "\n";
        echo "Reasoning: " . $result->detection->reasoning . "\n";

        if ($result->details) {
            echo "Terms: " . implode(', ', $result->details->dogwhistleTerms) . "\n";
            echo "Categories: " . implode(', ', $result->details->categories) . "\n";
            echo "Subtlety: " . $result->details->subtletyLevel . "\n";
            echo "Harm potential: " . $result->details->harmPotential . "\n";
        }
    });

$client->run();

Try it out live →

curl -X POST https://app.respectify.ai/v0.2/dogwhistle \
  -H "X-User-Email: your@email.com" \
  -H "X-API-Key: your-api-key" \
  -d "article_context_id=550e8400-e29b-41d4-a716-446655440000" \
  -d "comment=This comment to analyze for coded language"

Response:

{
  "detection": {
    "reasoning": "No coded language or dogwhistles detected in this comment.",
    "dogwhistles_detected": false,
    "confidence": 0.92
  }
}

With Custom Detection

Provide context for better detection:

Python
PHP
REST API

Blocking Client
Async Client

result = client.check_dogwhistle(
    "Comment text here",
    article_id,
    sensitive_topics=['topic-area-1', 'topic-area-2'],
    dogwhistle_examples=['known-coded-phrase', 'another-term']
)

result = await client.check_dogwhistle(
    "Comment text here",
    article_id,
    sensitive_topics=['topic-area-1', 'topic-area-2'],
    dogwhistle_examples=['known-coded-phrase', 'another-term']
)

$client->checkDogwhistle(
    $articleId,
    "Comment text here",
    ['topic-area-1', 'topic-area-2'],           // sensitive topics
    ['known-coded-phrase', 'another-term']      // known examples
);

Try it out live →

curl -X POST https://app.respectify.ai/v0.2/dogwhistle \
  -H "X-User-Email: your@email.com" \
  -H "X-API-Key: your-api-key" \
  -d "article_context_id=550e8400-e29b-41d4-a716-446655440000" \
  -d "comment=Comment text here" \
  -d "sensitive_topics[]=topic-area-1" \
  -d "sensitive_topics[]=topic-area-2" \
  -d "dogwhistle_examples[]=known-coded-phrase" \
  -d "dogwhistle_examples[]=another-term"

Reference​

By Example​

Basic Usage​

With Custom Detection​

Reference

By Example

Basic Usage

With Custom Detection