Skip to main content

​
Ollama Service API

The MOOD MNKY Ollama service provides enterprise-grade AI model management and inference capabilities. This documentation covers all available endpoints, authentication methods, and best practices for integration.

​
Base URL

https://ollama.moodmnky.com

​
Available Endpoints

​
Model Management

​
Generation & Inference

​
Monitoring & Health

​
Authentication

All requests to the Ollama service must include an API key in the Authorization header: 
Authorization: Bearer your-api-key
To obtain an API key:
  1. Contact the DevOps team through the developer portal
  2. Specify your use case and required rate limits
  3. Follow our security best practices for API key management

​
Request Format

All POST requests should use JSON with the content type header: 
Content-Type: application/json

​
Response Format

Successful responses will have appropriate HTTP status codes and JSON bodies: 
{
  "status": "success",
  "data": {
    // Response data here
  }
}
Error responses follow the standard error format: 
{
  "error": {
    "code": "error_code",
    "message": "Error description",
    "details": {
      // Additional error details
    }
  }
}

​
Rate Limiting

The service implements the following rate limits:
Endpoint CategoryRate LimitBurst Limit
Generation100/min120/min
Model Management1000/min1200/min
Monitoring1000/min1200/min
Rate limit headers are included in all responses:
  • X-RateLimit-Limit: Rate limit ceiling
  • X-RateLimit-Remaining: Remaining requests
  • X-RateLimit-Reset: Time until limit reset

​
Monitoring

The service exposes Prometheus metrics at: 
https://ollama.moodmnky.com/metrics
Available metrics include:
  • Request counts and latencies
  • Model loading/unloading events
  • Resource utilization
  • Error rates
  • Token usage

​
Best Practices

  1. Model Management
    • Cache model information locally
    • Implement exponential backoff for retries
    • Monitor model versions
  2. Generation Requests
    • Use streaming for long generations
    • Implement request timeouts
    • Handle rate limits gracefully
  3. Production Usage
    • Monitor API response times
    • Set up alerts for error rates
    • Track token usage
    • Implement circuit breakers

​
Examples

​
Curl

# Generate a completion
curl -X POST https://ollama.moodmnky.com/api/generate \
  -H "Authorization: Bearer your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama3.2",
    "prompt": "Why is the sky blue?"
  }'

​
Python

import requests

API_KEY = "your-api-key"
BASE_URL = "https://ollama.moodmnky.com"

def generate_completion(prompt, model="llama3.2"):
    response = requests.post(
        f"{BASE_URL}/api/generate",
        headers={
            "Authorization": f"Bearer {API_KEY}",
            "Content-Type": "application/json"
        },
        json={
            "model": model,
            "prompt": prompt
        }
    )
    return response.json()

​
JavaScript

const API_KEY = 'your-api-key';
const BASE_URL = 'https://ollama.moodmnky.com';

async function generateCompletion(prompt, model = 'llama3.2') {
  const response = await fetch(`${BASE_URL}/api/generate`, {
    method: 'POST',
    headers: {
      'Authorization': `Bearer ${API_KEY}`,
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({
      model,
      prompt
    })
  });
  return await response.json();
}

​
Support