curl -X POST "https://api.tensorone.ai/v2/ai/uncensored-chat" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {
        "role": "system",
        "content": "You are a creative writing assistant helping with fiction."
      },
      {
        "role": "user",
        "content": "Write a dark, gritty detective story opening with moral ambiguity"
      }
    ],
    "model": "uncensored-llama-70b",
    "maxTokens": 800,
    "temperature": 0.9,
    "safetyLevel": "minimal"
  }'
{
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "Detective Maria Santos stood in the rain-soaked alley, staring at the body sprawled between overflowing dumpsters. Third one this month. The victim—a known informant who'd been feeding her information about the Kozlov organization—had taken two bullets to the chest.\n\nShe'd told him to lay low. Hell, she'd practically begged him. But Jimmy needed the money, needed it bad enough to risk everything. Now his kids would grow up without a father, and the case she'd been building for two years might die with him.\n\nMaria lit a cigarette with shaking hands. The smart play was to call it in, let the uniforms handle the scene, file her report. But Jimmy's last text haunted her: 'Meeting tomorrow. Big fish. Location attached.' \n\nShe looked at her phone. The address was still there. A warehouse in the industrial district. She could wait for backup, follow protocol, let Kozlov slip through the cracks again. Or she could go alone, break every rule in the book, and maybe—just maybe—get justice for Jimmy and a dozen other victims.\n\nThe rain kept falling. Maria flicked away her cigarette and made her choice."
      },
      "safetyFlags": []
    }
  ],
  "usage": {
    "promptTokens": 58,
    "completionTokens": 187,
    "totalTokens": 245
  }
}
This endpoint provides access to AI models without standard content filtering. It is intended for research, creative writing, and development purposes only. Users must comply with applicable laws and TensorOne’s terms of service.
The Uncensored AI Chat endpoint provides access to language models that have minimal content restrictions, enabling more open-ended conversations and creative applications while maintaining safety through usage monitoring.

Request Body

messages
array
required
Array of message objects representing the conversation history
model
string
default:"uncensored-llama-70b"
Uncensored model to use:
  • uncensored-llama-70b - Large model with minimal filtering
  • uncensored-mistral-7b - Smaller, faster model
  • uncensored-mixtral-8x7b - Mixture of experts model
maxTokens
integer
default:"2000"
Maximum number of tokens to generate
temperature
number
default:"0.8"
Creativity level (0.0 to 1.5). Higher values increase randomness.
safetyLevel
string
default:"minimal"
Safety filtering level:
  • none - No content filtering
  • minimal - Basic harmful content detection only
  • moderate - Standard content filtering (defeats the purpose)

Response

choices
array
Array of completion choices
usage
object
Token usage and monitoring information

Example

curl -X POST "https://api.tensorone.ai/v2/ai/uncensored-chat" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {
        "role": "system",
        "content": "You are a creative writing assistant helping with fiction."
      },
      {
        "role": "user",
        "content": "Write a dark, gritty detective story opening with moral ambiguity"
      }
    ],
    "model": "uncensored-llama-70b",
    "maxTokens": 800,
    "temperature": 0.9,
    "safetyLevel": "minimal"
  }'
{
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "Detective Maria Santos stood in the rain-soaked alley, staring at the body sprawled between overflowing dumpsters. Third one this month. The victim—a known informant who'd been feeding her information about the Kozlov organization—had taken two bullets to the chest.\n\nShe'd told him to lay low. Hell, she'd practically begged him. But Jimmy needed the money, needed it bad enough to risk everything. Now his kids would grow up without a father, and the case she'd been building for two years might die with him.\n\nMaria lit a cigarette with shaking hands. The smart play was to call it in, let the uniforms handle the scene, file her report. But Jimmy's last text haunted her: 'Meeting tomorrow. Big fish. Location attached.' \n\nShe looked at her phone. The address was still there. A warehouse in the industrial district. She could wait for backup, follow protocol, let Kozlov slip through the cracks again. Or she could go alone, break every rule in the book, and maybe—just maybe—get justice for Jimmy and a dozen other victims.\n\nThe rain kept falling. Maria flicked away her cigarette and made her choice."
      },
      "safetyFlags": []
    }
  ],
  "usage": {
    "promptTokens": 58,
    "completionTokens": 187,
    "totalTokens": 245
  }
}

Use Cases

Creative Writing

  • Fiction with mature themes
  • Complex character development
  • Morally ambiguous scenarios
  • Historical fiction with accurate context

Research & Education

  • Academic discussions on sensitive topics
  • Philosophical debates
  • Historical analysis without sanitization
  • Medical/scientific content

Game Development

  • NPC dialogue for mature games
  • Branching narrative content
  • World-building with realistic conflicts
  • Interactive storytelling

Content Creation

  • Screenplay writing
  • Novel development
  • Documentary research
  • Investigative journalism

Safety & Compliance

While this endpoint has minimal content filtering, all conversations are logged for safety monitoring. Violations of terms of service may result in account suspension.

Prohibited Uses

  • Generating illegal content
  • Creating harmful instructions
  • Harassment or hate speech
  • Impersonating real individuals maliciously
  • Creating misleading information as fact

Best Practices

  • Clearly mark generated content as AI-created
  • Use appropriate context in system messages
  • Implement your own content filters for public applications
  • Regular monitoring of generated content
  • Respect platform terms and local laws

Rate Limits

  • Research Tier: 100 requests/hour
  • Professional Tier: 500 requests/hour
  • Enterprise Tier: Custom limits available
Higher rate limits require verification of legitimate use case and compliance measures.

Monitoring

All uncensored chat requests are subject to:
  • Automated safety monitoring
  • Usage pattern analysis
  • Regular compliance reviews
  • Potential human review for policy violations
Accounts showing concerning patterns may be temporarily restricted pending review.