🤖 Connect ANY AI to Your App — The Complete Beginner's Guide

One guide. Every AI provider. Zero confusion. Works with Android, iOS, Web, Desktop — any app that can make HTTP requests.

---

📋 Table of Contents

1. What is an API Key?
2. Google Gemini (FREE)
3. OpenAI / ChatGPT
4. Anthropic Claude
5. OpenRouter (400+ Models)
6. Groq (Ultra Fast)
7. Ollama (Local, Free Forever)
8. How API Calls Work (Visual)
9. Code Examples
10. Multi-Key Failover
11. Troubleshooting
12. Security Tips

---

🔑 What is an API Key?

Think of an API key like a password for your app to talk to an AI service.

You (App)  ──── API Key ────►  AI Service (Gemini, ChatGPT, etc.)
                                    │
                                    ▼
                              AI Response

It's just a long string of characters, like:

• Gemini: AIzaSyD_xxxxxxxxxxxxxxxxxxxx
• OpenAI: sk-proj-xxxxxxxxxxxxxxxxxxxx
• Anthropic: sk-ant-api03-xxxxxxxxxxxx

You get it for free (most providers have free tiers), paste it into your app, and boom — your app can talk to AI.

---

1. 🔵 Google Gemini (FREE — Easiest!)

Why Start Here?

• ✅ Completely FREE (generous limits)
• ✅ Easiest setup (2 minutes)
• ✅ Powerful models (Gemini 2.0 Flash, 2.5 Pro)
• ✅ Text + Image + Audio + Video support

Step-by-Step: Get Your Key

Step 1: Open browser → Go to https://aistudio.google.com
Step 2: Sign in with your Google account
Step 3: Click "Get API Key" (top-left or sidebar)
Step 4: Click "Create API Key"
Step 5: Copy the key (starts with "AIzaSy...")
Step 6: Paste it in your app → Done! 🎉

Available Models

Model	Speed	Quality	Best For
gemini-2.0-flash	⚡⚡⚡	★★★★	General use (RECOMMENDED)
gemini-2.0-flash-lite	⚡⚡⚡⚡	★★★	Quick simple tasks
gemini-1.5-flash	⚡⚡⚡	★★★★	Balanced
gemini-1.5-pro	⚡⚡	★★★★★	Complex reasoning
gemini-2.5-flash-preview-05-20	⚡⚡⚡	★★★★★	Latest & greatest
gemini-2.5-pro-preview-05-06	⚡	★★★★★+	Most powerful

API Call Format

POST https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash:generateContent?key=YOUR_KEY

Headers:
  Content-Type: application/json

Body:
{
  "contents": [
    {
      "parts": [
        { "text": "Your question here" }
      ]
    }
  ],
  "generationConfig": {
    "temperature": 0.7,
    "maxOutputTokens": 4096
  }
}

Read the Response

{
  "candidates": [
    {
      "content": {
        "parts": [
          { "text": "The AI's answer is here!" }
        ]
      }
    }
  ]
}

Get the answer: response.candidates[0].content.parts[0].text

---

2. 🟢 OpenAI / ChatGPT

Get Your Key

Step 1: Go to https://platform.openai.com
Step 2: Sign up / Log in
Step 3: Click your profile icon → "API Keys"
       (or go to https://platform.openai.com/api-keys)
Step 4: Click "Create new secret key"
Step 5: Name it anything (e.g., "My App")
Step 6: Copy the key (starts with "sk-proj-...")
Step 7: ⚠️ SAVE IT NOW — you can't see it again!

💰 Pricing: Free $5 credits for new accounts. After that, pay-as-you-go. GPT-4o-mini is very cheap (~$0.15 per 1M tokens).

Available Models

Model	Speed	Quality	Cost
gpt-4o-mini	⚡⚡⚡	★★★★	💲 (Cheapest)
gpt-4o	⚡⚡	★★★★★	💲💲
gpt-4-turbo	⚡⚡	★★★★★	💲💲💲
o3-mini	⚡	★★★★★+	💲💲💲 (Reasoning)

API Call Format

POST https://api.openai.com/v1/chat/completions

Headers:
  Content-Type: application/json
  Authorization: Bearer YOUR_API_KEY

Body:
{
  "model": "gpt-4o-mini",
  "messages": [
    { "role": "system", "content": "You are a helpful assistant." },
    { "role": "user", "content": "Your question here" }
  ],
  "max_tokens": 1000
}

Read the Response

{
  "choices": [
    {
      "message": {
        "content": "The AI's answer is here!"
      }
    }
  ]
}

Get the answer: response.choices[0].message.content

---

3. 🟠 Anthropic Claude

Get Your Key

Step 1: Go to https://console.anthropic.com
Step 2: Sign up (may need to add payment method)
Step 3: Go to "API Keys" in the sidebar
Step 4: Click "Create Key"
Step 5: Copy the key (starts with "sk-ant-api03-...")

💰 Pricing: Free $5 credits for new accounts. Claude Haiku is cheapest.

Available Models

Model	Speed	Quality	Cost
claude-haiku-4-5-20251001	⚡⚡⚡	★★★★	💲 (Cheapest)
claude-sonnet-4-6	⚡⚡	★★★★★	💲💲
claude-opus-4-6	⚡	★★★★★+	💲💲💲 (Best)

API Call Format

POST https://api.anthropic.com/v1/messages

Headers:
  Content-Type: application/json
  x-api-key: YOUR_API_KEY
  anthropic-version: 2023-06-01

Body:
{
  "model": "claude-haiku-4-5-20251001",
  "max_tokens": 1024,
  "system": "You are a helpful assistant.",
  "messages": [
    { "role": "user", "content": "Your question here" }
  ]
}

Read the Response

{
  "content": [
    {
      "type": "text",
      "text": "The AI's answer is here!"
    }
  ]
}

Get the answer: response.content[0].text

---

4. 🟣 OpenRouter (400+ Models — One Key, ALL AIs!)

Why OpenRouter?

• ✅ ONE API key → access OpenAI, Claude, Gemini, Llama, Mistral, and 400+ models
• ✅ Some FREE models available
• ✅ Same OpenAI-compatible format for ALL models
• ✅ Pay only for what you use

Get Your Key

Step 1: Go to https://openrouter.ai
Step 2: Sign up with Google/GitHub
Step 3: Go to https://openrouter.ai/keys
Step 4: Click "Create Key"
Step 5: Copy the key (starts with "sk-or-v1-...")

Popular Models on OpenRouter

Model	Provider	Cost
google/gemini-2.0-flash-exp:free	Google	FREE
meta-llama/llama-3.1-8b-instruct:free	Meta	FREE
openai/gpt-4o-mini	OpenAI	💲
anthropic/claude-sonnet-4-6	Anthropic	💲💲
google/gemini-2.5-pro-preview	Google	💲💲

API Call Format (Same as OpenAI!)

POST https://openrouter.ai/api/v1/chat/completions

Headers:
  Content-Type: application/json
  Authorization: Bearer YOUR_OPENROUTER_KEY
  HTTP-Referer: https://your-app-name.com

Body:
{
  "model": "google/gemini-2.0-flash-exp:free",
  "messages": [
    { "role": "user", "content": "Your question here" }
  ]
}

Read the Response (Same as OpenAI!)

{
  "choices": [
    {
      "message": {
        "content": "The AI's answer is here!"
      }
    }
  ]
}

Get the answer: response.choices[0].message.content

---

5. ⚡ Groq (Ultra Fast — Free Tier)

Why Groq?

• ✅ INSANELY fast (10x faster than others)
• ✅ Free tier available
• ✅ Runs Llama, Mixtral, Gemma models
• ✅ OpenAI-compatible API

Get Your Key

Step 1: Go to https://console.groq.com
Step 2: Sign up with Google/GitHub
Step 3: Go to "API Keys"
Step 4: Click "Create API Key"
Step 5: Copy the key (starts with "gsk_...")

Available Models

Model	Speed	Best For
llama-3.1-70b-versatile	⚡⚡⚡⚡	General use
llama-3.1-8b-instant	⚡⚡⚡⚡⚡	Quick tasks
mixtral-8x7b-32768	⚡⚡⚡⚡	Long context
gemma2-9b-it	⚡⚡⚡⚡	Lightweight

API Call Format (OpenAI-Compatible!)

POST https://api.groq.com/openai/v1/chat/completions

Headers:
  Content-Type: application/json
  Authorization: Bearer YOUR_GROQ_KEY

Body:
{
  "model": "llama-3.1-70b-versatile",
  "messages": [
    { "role": "user", "content": "Your question here" }
  ],
  "max_tokens": 1000
}

Response format: Same as OpenAI → response.choices[0].message.content

---

6. ⚫ Ollama (100% FREE, Runs on Your PC)

Why Ollama?

• ✅ Completely FREE forever — no API key, no limits
• ✅ Runs on your own computer
• ✅ Your data never leaves your machine
• ✅ Hundreds of open-source models

Setup

Step 1: Go to https://ollama.com
Step 2: Download & install for your OS
Step 3: Open terminal/command prompt
Step 4: Run: ollama pull mistral    (downloads a model)
Step 5: It's running! No key needed.

Popular Models

ollama pull mistral          # 7B - Great general purpose
ollama pull llama3.1         # 8B - Meta's best open model
ollama pull codellama        # 7B - Coding specialist
ollama pull gemma2           # 9B - Google's open model
ollama pull phi3             # 3.8B - Microsoft, super small

API Call Format

POST http://127.0.0.1:11434/api/chat

Headers:
  Content-Type: application/json

Body:
{
  "model": "mistral",
  "messages": [
    { "role": "user", "content": "Your question here" }
  ],
  "stream": false
}

No API key needed! Just make sure Ollama is running on your computer.

Read the Response

{
  "message": {
    "content": "The AI's answer is here!"
  }
}

Get the answer: response.message.content

---

🔄 How API Calls Actually Work

┌─────────────┐         ┌──────────────────┐         ┌────────────┐
│             │  HTTPS   │                  │  Magic   │            │
│  YOUR APP   │────────► │   AI Provider    │────────► │  AI Model  │
│             │          │  (Google, etc.)  │          │            │
│             │ ◄─────── │                  │ ◄─────── │            │
│             │  JSON    │                  │  Result  │            │
└─────────────┘         └──────────────────┘         └────────────┘

What happens:
1. Your app sends a POST request with your question + API key
2. The AI provider validates your key
3. The AI model processes your question
4. The provider sends back a JSON response
5. Your app reads the answer from the JSON

The Universal Pattern (ALL providers follow this):

1. URL      → Where to send the request
2. Headers  → Your API key + content type
3. Body     → Your question (JSON format)
4. Response → AI's answer (JSON format)

---

📦 Copy-Paste Code Examples

JavaScript (Web / Node.js)

// 🔵 GEMINI
async function askGemini(question, apiKey) {
  const res = await fetch(
    `https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash:generateContent?key=${apiKey}`,
    {
      method: 'POST',
      headers: { 'Content-Type': 'application/json' },
      body: JSON.stringify({
        contents: [{ parts: [{ text: question }] }],
        generationConfig: { temperature: 0.7, maxOutputTokens: 4096 }
      })
    }
  );
  const data = await res.json();
  return data.candidates?.[0]?.content?.parts?.[0]?.text || 'No response';
}

// 🟢 OPENAI / 🟣 OPENROUTER / ⚡ GROQ (all same format!)
async function askOpenAI(question, apiKey, baseUrl = 'https://api.openai.com', model = 'gpt-4o-mini') {
  const res = await fetch(`${baseUrl}/v1/chat/completions`, {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'Authorization': `Bearer ${apiKey}`
    },
    body: JSON.stringify({
      model: model,
      messages: [{ role: 'user', content: question }],
      max_tokens: 1000
    })
  });
  const data = await res.json();
  return data.choices?.[0]?.message?.content || 'No response';
}

// 🟠 ANTHROPIC
async function askClaude(question, apiKey, model = 'claude-haiku-4-5-20251001') {
  const res = await fetch('https://api.anthropic.com/v1/messages', {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'x-api-key': apiKey,
      'anthropic-version': '2023-06-01'
    },
    body: JSON.stringify({
      model: model,
      max_tokens: 1024,
      messages: [{ role: 'user', content: question }]
    })
  });
  const data = await res.json();
  return data.content?.[0]?.text || 'No response';
}

// Usage:
// const answer = await askGemini("What is Kotlin?", "AIzaSy...");
// const answer = await askOpenAI("What is Kotlin?", "sk-proj-...");
// const answer = await askOpenAI("What is Kotlin?", "sk-or-v1-...", "https://openrouter.ai/api", "google/gemini-2.0-flash-exp:free");
// const answer = await askOpenAI("What is Kotlin?", "gsk_...", "https://api.groq.com/openai", "llama-3.1-70b-versatile");
// const answer = await askClaude("What is Kotlin?", "sk-ant-api03-...");

Kotlin (Android)

// Using OkHttp (add to build.gradle: implementation("com.squareup.okhttp3:okhttp:4.12.0"))

import okhttp3.*
import okhttp3.MediaType.Companion.toMediaType
import okhttp3.RequestBody.Companion.toRequestBody
import org.json.JSONObject

// 🔵 GEMINI
suspend fun askGemini(question: String, apiKey: String, model: String = "gemini-2.0-flash"): String {
    val client = OkHttpClient()
    val json = JSONObject().apply {
        put("contents", org.json.JSONArray().put(
            JSONObject().put("parts", org.json.JSONArray().put(
                JSONObject().put("text", question)
            ))
        ))
        put("generationConfig", JSONObject().apply {
            put("temperature", 0.7)
            put("maxOutputTokens", 4096)
        })
    }
    val request = Request.Builder()
        .url("https://generativelanguage.googleapis.com/v1beta/models/$model:generateContent?key=$apiKey")
        .post(json.toString().toRequestBody("application/json".toMediaType()))
        .build()
    val response = client.newCall(request).execute()
    val body = JSONObject(response.body!!.string())
    return body.getJSONArray("candidates")
        .getJSONObject(0).getJSONObject("content")
        .getJSONArray("parts").getJSONObject(0)
        .getString("text")
}

// 🟢 OPENAI-COMPATIBLE (works for OpenAI, OpenRouter, Groq)
suspend fun askOpenAI(
    question: String,
    apiKey: String,
    baseUrl: String = "https://api.openai.com",
    model: String = "gpt-4o-mini"
): String {
    val client = OkHttpClient()
    val json = JSONObject().apply {
        put("model", model)
        put("messages", org.json.JSONArray().put(
            JSONObject().put("role", "user").put("content", question)
        ))
        put("max_tokens", 1000)
    }
    val request = Request.Builder()
        .url("$baseUrl/v1/chat/completions")
        .addHeader("Authorization", "Bearer $apiKey")
        .post(json.toString().toRequestBody("application/json".toMediaType()))
        .build()
    val response = client.newCall(request).execute()
    val body = JSONObject(response.body!!.string())
    return body.getJSONArray("choices")
        .getJSONObject(0).getJSONObject("message")
        .getString("content")
}

// 🟠 ANTHROPIC
suspend fun askClaude(question: String, apiKey: String, model: String = "claude-haiku-4-5-20251001"): String {
    val client = OkHttpClient()
    val json = JSONObject().apply {
        put("model", model)
        put("max_tokens", 1024)
        put("messages", org.json.JSONArray().put(
            JSONObject().put("role", "user").put("content", question)
        ))
    }
    val request = Request.Builder()
        .url("https://api.anthropic.com/v1/messages")
        .addHeader("x-api-key", apiKey)
        .addHeader("anthropic-version", "2023-06-01")
        .post(json.toString().toRequestBody("application/json".toMediaType()))
        .build()
    val response = client.newCall(request).execute()
    val body = JSONObject(response.body!!.string())
    return body.getJSONArray("content")
        .getJSONObject(0).getString("text")
}

Python

import requests

# 🔵 GEMINI
def ask_gemini(question, api_key, model="gemini-2.0-flash"):
    url = f"https://generativelanguage.googleapis.com/v1beta/models/{model}:generateContent?key={api_key}"
    data = {
        "contents": [{"parts": [{"text": question}]}],
        "generationConfig": {"temperature": 0.7, "maxOutputTokens": 4096}
    }
    r = requests.post(url, json=data)
    return r.json()["candidates"][0]["content"]["parts"][0]["text"]

# 🟢 OPENAI-COMPATIBLE (OpenAI, OpenRouter, Groq)
def ask_openai(question, api_key, base_url="https://api.openai.com", model="gpt-4o-mini"):
    url = f"{base_url}/v1/chat/completions"
    headers = {"Authorization": f"Bearer {api_key}"}
    data = {
        "model": model,
        "messages": [{"role": "user", "content": question}],
        "max_tokens": 1000
    }
    r = requests.post(url, json=data, headers=headers)
    return r.json()["choices"][0]["message"]["content"]

# 🟠 ANTHROPIC
def ask_claude(question, api_key, model="claude-haiku-4-5-20251001"):
    url = "https://api.anthropic.com/v1/messages"
    headers = {"x-api-key": api_key, "anthropic-version": "2023-06-01"}
    data = {
        "model": model,
        "max_tokens": 1024,
        "messages": [{"role": "user", "content": question}]
    }
    r = requests.post(url, json=data, headers=headers)
    return r.json()["content"][0]["text"]

---

🔄 Pro Tip: Multi-Key Failover

If one key hits a rate limit, automatically switch to the next one:

const keys = [
  { provider: 'gemini', key: 'AIzaSy...', model: 'gemini-2.0-flash' },
  { provider: 'openrouter', key: 'sk-or-v1-...', model: 'google/gemini-2.0-flash-exp:free' },
  { provider: 'groq', key: 'gsk_...', model: 'llama-3.1-70b-versatile' },
];

async function askAI(question) {
  for (const entry of keys) {
    try {
      if (entry.provider === 'gemini') {
        return await askGemini(question, entry.key, entry.model);
      } else if (entry.provider === 'openrouter' || entry.provider === 'groq') {
        const baseUrl = entry.provider === 'openrouter'
          ? 'https://openrouter.ai/api'
          : 'https://api.groq.com/openai';
        return await askOpenAI(question, entry.key, baseUrl, entry.model);
      }
    } catch (err) {
      console.log(`${entry.provider} failed, trying next...`);
    }
  }
  throw new Error('All AI providers failed!');
}

---

❓ Troubleshooting

Error	What It Means	Fix
401 Unauthorized	Wrong API key	Double-check your key, no extra spaces
403 Forbidden	Key doesn't have access	Check billing/permissions on provider dashboard
429 Too Many Requests	Rate limited	Wait a minute, or use multi-key failover
500 Internal Server Error	Provider is down	Wait and retry, or switch provider
CORS Error (browser)	Browser blocking request	Use a backend proxy, or use the API from server-side
Model not found	Wrong model name	Check the model table above for exact names

Quick Debug Checklist

□ Is the API key correct? (no extra spaces, full key)
□ Is the API key enabled? (not revoked on dashboard)
□ Do you have credits/billing set up?
□ Is the model name spelled correctly?
□ Are you using the right URL for the provider?
□ Is the Content-Type header set to application/json?
□ Is the request body valid JSON?

---

🔒 Security Tips

⚠️  NEVER put API keys in:
    ✗ Source code committed to GitHub
    ✗ Client-side JavaScript (anyone can see it)
    ✗ Public URLs or logs

✅  ALWAYS store API keys in:
    ✓ Environment variables (.env files)
    ✓ Android SharedPreferences / DataStore
    ✓ iOS Keychain
    ✓ Server-side only (proxy API calls)
    ✓ localStorage (for personal/prototype apps only)

---

🎯 Quick Decision Chart

Need FREE AI?
├── Yes, with internet    → Gemini (best free) or OpenRouter (free models)
├── Yes, no internet      → Ollama (runs locally)
└── Budget available?
    ├── Want the BEST      → Claude Opus or GPT-4o
    ├── Want FAST + CHEAP  → Groq (Llama) or GPT-4o-mini
    ├── Want ONE key for ALL → OpenRouter
    └── Want PRIVACY       → Ollama (your data stays local)

---

🔗 Quick Links

Provider	Get API Key	Docs	Pricing
Gemini	aistudio.google.com	Docs	FREE
OpenAI	platform.openai.com	Docs	Pricing
Anthropic	console.anthropic.com	Docs	Pricing
OpenRouter	openrouter.ai/keys	Docs	Per model
Groq	console.groq.com	Docs	FREE tier
Ollama	ollama.com	Docs	FREE forever

---

Made by @ashokvarmamatta — If this helped you, give it a ⭐!

Used in: Prompt Gen · ZeroClaw · Neural Forge