API Documentation

Complete API reference for B01-NUna. All endpoints support JSON responses.

Loading API documentation...

Quick Start

Base URL

https://helloblue.ai

Authentication

Use either Bearer token or X-API-Key header:

Authorization: Bearer YOUR_API_KEYX-API-Key: YOUR_API_KEY

Endpoints

POST/api/v1/generate

Generate AI text completions using B01-NUna models.

Request Body:

{
  "model": "b01-1.2v-5b",
  "prompt": "Your prompt here",
  "max_tokens": 512,
  "temperature": 0.7
}

Response:

{
  "model": "b01-1.2v-5b",
  "text": "Generated response...",
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 366,
    "total_tokens": 378
  },
  "metadata": {
    "provider": "Groq",
    "processing_time": 581,
    "confidence": 0.85
  }
}

POST/api/chat

Interactive chat endpoint with streaming support, file attachments, and multi-modal capabilities.

Request Body (JSON):

{
  "message": "Hello!",
  "sessionId": "optional-session-id",
  "stream": true,
  "useReasoning": false,
  "useSearch": false,
  "useNeuroEngine": false
}

Request (Multipart/Form-Data):

For file uploads, use multipart/form-data:

FormData:
  - message: "Analyze this image"
  - sessionId: "optional-id"
  - file: [File object]
  - fileContext: [JSON string]

Response:

Server-Sent Events (SSE) stream:

data: {"type":"message","content":"Hello"}
data: {"type":"message","content":"!"}
data: {"type":"done"}

GET/api/health

System health check endpoint. No authentication required.

Response:

{
  "status": "healthy",
  "timestamp": "2026-01-15T10:30:00.000Z",
  "uptime": {
    "seconds": 3600,
    "formatted": "1h 0m 0s"
  },
  "memory": {
    "used": 150,
    "total": 512,
    "percentage": 29
  },
  "providers": {
    "total": 4,
    "available": 4,
    "healthy": 4
  }
}

POST/api/health

Trigger performance optimization. Rate limited.

Response:

{
  "status": "success",
  "message": "Optimization completed successfully",
  "results": {
    "memoryFreed": "25.50MB",
    "actions": ["cache_cleared", "gc_triggered"]
  }
}

GET/api/training

Get training status, GPU information, or data statistics.

Query Parameters:

  • action=status - Training status and recommendations (default)
  • action=gpu-status - GPU availability and specifications
  • action=data-stats - Training data statistics
  • action=checkpoint-info - Checkpoint information for a model
  • domain - Optional domain filter (coding, creative, analytical, general)
  • model - Model name for checkpoint-info action

Example Response (status):

{
  "success": true,
  "data": {
    "decision": {
      "shouldTrain": false,
      "reason": "System performance is good",
      "priority": "low"
    },
    "statistics": {
      "totalExamples": 180,
      "averageQuality": 0.831
    },
    "trainingInProgress": false
  }
}

POST/api/training

Start GPU-based model training. Training runs asynchronously in the background.

Request Body:

{
  "domain": "coding",
  "force": false,
  "resume": false
}

Response:

{
  "success": true,
  "message": "Training started",
  "domain": "coding"
}

POST/api/training/test-data

Add test training data for development and testing purposes.

Request Body:

{
  "count": 60
}

OpenAPI Specification

Full OpenAPI 3.0 specification is available for integration with API tools.

View OpenAPI Spec (JSON)Open in Swagger EditorOpen in Swagger UI

API Provider: Helloblue AI
Support: support@helloblue.ai

Available Models

B01-NUna

Primary conversational AI model with multi-modal capabilities.

helloblueai/B01-NUna

Bleu XGBoost Classifier

Advanced XGBoost-based classifier for text classification and quality assessment. Integrated with B01-NUna for intelligent routing and classification tasks.

helloblueai/bleu-xgboost-classifierView on Hugging Face

Error Codes

400Bad Request - Invalid request format
401Unauthorized - Missing or invalid API key
500Internal Server Error - Server-side error