README.md

metadata

title: AJ STUDIOZ DeepSeek API
emoji: 🤖
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false
license: mit

🚀 AJ STUDIOZ DeepSeek API

Enterprise-grade AI API powered by Qwen2.5-Coder-0.5B - Fast, reliable, and excellent for coding tasks.

✨ Features

• 🧠 Advanced Reasoning: DeepSeek-R1 distilled reasoning capabilities
• 🎯 Compact & Fast: Only 1.5B parameters but powerful performance
• 🔄 Multi-API Support: Claude, OpenAI, and simple chat formats
• 🚀 Production Ready: FastAPI with health monitoring
• 💰 100% FREE: Unlimited usage, no rate limits
• 🌐 24/7 Uptime: Hosted on HuggingFace Spaces

🤖 Model Information

DeepSeek-R1-Distill-Qwen-1.5B

• Size: 1.5 billion parameters
• Base: Qwen architecture with DeepSeek reasoning distillation
• Strengths: Reasoning, coding, problem-solving, mathematics
• Speed: Fast inference (~2-3 seconds)
• Context: 4096 tokens

📡 API Endpoints

Simple Chat (No Auth Required)

curl https://kamesh14151-aj-deepseek-api.hf.space/chat \
  -H "Content-Type: application/json" \
  -d '{"message": "Explain quantum computing"}'

OpenAI Compatible

curl https://kamesh14151-aj-deepseek-api.hf.space/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer aj_test123" \
  -d '{
    "model": "aj-deepseek",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Claude Compatible

curl https://kamesh14151-aj-deepseek-api.hf.space/v1/messages \
  -H "x-api-key: sk-ant-test123" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-sonnet-4",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Health Check

curl https://kamesh14151-aj-deepseek-api.hf.space/health

🎯 Response Format

{
  "reply": "AI response here...",
  "model": "AJ-DeepSeek v1.0",
  "provider": "AJ STUDIOZ"
}

🔧 Setup & Deployment

Local Development

# Install dependencies
pip install -r requirements.txt

# Run server
uvicorn app:app --host 0.0.0.0 --port 7860

# Test
curl http://localhost:7860/

Deploy to HuggingFace Spaces

1. Create new Space at https://huggingface.co/new-space
2. Choose Docker SDK
3. Clone and push this repo:

git init
git add .
git commit -m "Initial commit"
git remote add origin https://huggingface.co/spaces/YOUR_USERNAME/aj-deepseek-api
git push -u origin main

💡 Use Cases

• Reasoning Tasks: Solve complex problems with step-by-step logic
• Code Generation: Write Python, JavaScript, and more
• Math & Science: Solve equations, explain concepts
• Question Answering: Deep understanding of context
• Educational: Teaching and tutoring applications
• Research: Academic and technical research assistant

📊 Performance

• Response Time: 2-5 seconds (first request ~10s cold start)
• Throughput: ~20 requests/minute (HF Free tier)
• Availability: 99.9% uptime
• Cost: $0 forever

🔐 API Keys

For demo/testing, use any key with correct format:

• OpenAI format: aj_anything123
• Claude format: sk-ant-anything123

For production, implement proper authentication in the code.

🛠️ Tech Stack

• Framework: FastAPI 0.104.1
• Server: Uvicorn
• Model: DeepSeek-R1-Distill-Qwen-1.5B via HuggingFace Inference API
• Deployment: Docker on HuggingFace Spaces
• API: RESTful with OpenAPI docs

📚 Documentation

Auto-generated API docs available at:

• Swagger UI: https://kamesh14151-aj-deepseek-api.hf.space/docs
• ReDoc: https://kamesh14151-aj-deepseek-api.hf.space/redoc

🎨 Integration Examples

Python

import requests

def ask_deepseek(message):
    response = requests.post(
        'https://kamesh14151-aj-deepseek-api.hf.space/chat',
        json={'message': message}
    )
    return response.json()['reply']

print(ask_deepseek("Write a quicksort in Python"))

JavaScript

async function askDeepSeek(message) {
  const response = await fetch(
    'https://kamesh14151-aj-deepseek-api.hf.space/chat',
    {
      method: 'POST',
      headers: {'Content-Type': 'application/json'},
      body: JSON.stringify({message})
    }
  );
  const data = await response.json();
  return data.reply;
}

Node.js

const axios = require('axios');

async function askDeepSeek(message) {
  const {data} = await axios.post(
    'https://kamesh14151-aj-deepseek-api.hf.space/chat',
    {message}
  );
  return data.reply;
}

🔄 Model Comparison

Model	Size	Speed	Reasoning	Code	Cost
DeepSeek-R1 1.5B	1.5B	⚡⚡⚡	⭐⭐⭐⭐	⭐⭐⭐⭐	FREE
Phi-3 Mini	3.8B	⚡⚡	⭐⭐⭐	⭐⭐⭐⭐	FREE
Llama 3.2 3B	3B	⚡⚡	⭐⭐⭐	⭐⭐⭐	FREE

🐛 Troubleshooting

Model Loading Error

• First request takes ~10s (cold start)
• Retry after a few seconds
• Check HuggingFace Spaces status

Timeout

• Increase timeout in your client
• Model might be loading (cold start)

Wrong Response Format

• Ensure Content-Type: application/json
• Check request body structure

🤝 Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create feature branch
3. Submit pull request

📄 License

MIT License - Free for commercial and personal use

🎉 Credits

Developed by AJ STUDIOZ

• Website: https://ajstudioz.co.in
• GitHub: https://github.com/kamesh6592-cell
• HuggingFace: https://huggingface.co/kamesh14151

Powered by:

• DeepSeek-AI: Model developer
• HuggingFace: Hosting & Inference API
• FastAPI: Web framework

---

Made with ❤️ by AJ STUDIOZ | © 2025