Using the Gemini API with Python (Free Tier)

Build AI-powered applications using Google's Gemini model via the free tier.

Prerequisites

• Python 3.8+
• OpenRouter API key (free at https://openrouter.ai)
• requests library: pip install requests

Quick Start

import requests
import os

# Configuration
OPENROUTER_URL = "https://openrouter.ai/api/v1/chat/completions"
MODEL = "google/gemini-2.0-flash-exp:free"  # Free tier model

def call_gemini(prompt: str, api_key: str, max_tokens: int = 200) -> str:
    """Call the free Gemini API via OpenRouter."""
    headers = {
        "Authorization": f"Bearer {api_key}",
        "Content-Type": "application/json"
    }
    data = {
        "model": MODEL,
        "messages": [{"role": "user", "content": prompt}],
        "temperature": 0.7,
        "max_tokens": max_tokens
    }

    response = requests.post(OPENROUTER_URL, headers=headers, json=data)

    if response.status_code == 200:
        return response.json()["choices"][0]["message"]["content"]
    else:
        raise Exception(f"API Error: {response.status_code} - {response.text}")

# Usage
if __name__ == "__main__":
    api_key = os.getenv("OPENROUTER_API_KEY")
    if not api_key:
        print("Set OPENROUTER_API_KEY environment variable")
        exit(1)

    prompt = "Write a haiku about Python programming"
    try:
        result = call_gemini(prompt, api_key)
        print(f"Prompt: {prompt}")
        print(f"Response: {result}")
    except Exception as e:
        print(f"Error: {e}")

Advanced Example: Chat Loop

import requests
import os

def chat_with_gemini(api_key: str):
    """Interactive chat with Gemini."""
    conversation = []
    url = "https://openrouter.ai/api/v1/chat/completions"
    headers = {
        "Authorization": f"Bearer {api_key}",
        "Content-Type": "application/json"
    }

    print("Gemini Chat (type 'quit' to exit)")
    print("-" * 40)

    while True:
        user_input = input("You: ")
        if user_input.lower() in ('quit', 'exit', 'q'):
            break

        conversation.append({"role": "user", "content": user_input})

        data = {
            "model": "google/gemini-2.0-flash-exp:free",
            "messages": conversation,
            "temperature": 0.7,
            "max_tokens": 300
        }

        response = requests.post(url, headers=headers, json=data)

        if response.status_code == 200:
            assistant_msg = response.json()["choices"][0]["message"]["content"]
            conversation.append({"role": "assistant", "content": assistant_msg})
            print(f"Gemini: {assistant_msg}")
        else:
            print(f"Error: {response.status_code}")

        print("-" * 40)

if __name__ == "__main__":
    api_key = os.getenv("OPENROUTER_API_KEY")
    chat_with_gemini(api_key)

Key Parameters

Parameter	Range	Purpose
temperature	0.0-1.0	Randomness (low=factual, high=creative)
max_tokens	1-4096	Response length limit
model	various	Choose free models with :free suffix

Error Handling

import time

def robust_call(prompt: str, api_key: str, retries: int = 3) -> str:
    """Call API with exponential backoff retry."""
    for attempt in range(retries):
        try:
            return call_gemini(prompt, api_key)
        except Exception as e:
            if "429" in str(e):  # Rate limited
                wait = 2 ** attempt
                print(f"Rate limited. Waiting {wait}s...")
                time.sleep(wait)
            else:
                raise
    raise Exception("Max retries exceeded")

Best Practices

1. Secure API keys - Use environment variables
2. Handle rate limits - Implement exponential backoff
3. Limit token usage - Set max_tokens to control costs
4. Validate inputs - Sanitize user prompts
5. Log usage - Track API calls for debugging

Conclusion

The free Gemini tier via OpenRouter provides an excellent way to integrate AI into Python applications at zero cost. Start with the simple function above, then expand into chatbots, content generators, or analysis tools.