OpenAI API Key: How to Get One and Use It (2026 Guide)

Quick Answer

Visit platform.openai.com → sign up → add payment method → set usage limit → API Keys → Create new secret key → copy immediately.The key is shown once. Store in environment variables. ChatGPT Plus does NOT include API access — they're separate billing systems. New API accounts get $5 free credits valid for 3 months.

Step 1: Sign up at platform.openai.com

Go to platform.openai.com. The platform account is separate from chatgpt.com — you may need to create a new account or sign in with the same Google/Microsoft SSO and accept developer terms.

Step 2: Add a Payment Method

1. Click your profile icon → Settings → Billing → Payment Methods
2. Add a credit card
3. Optionally pre-pay credits via Add to Credit Balance

New accounts receive $5 in free credits valid for 3 months — enough to test most workloads.

Step 3: Set Usage Limits (Critical)

Before creating any API keys, set hard usage limits to prevent runaway charges:

1. Settings → Billing → Usage Limits
2. Set a monthly hard limit (e.g., $50)
3. Set a monthly soft limit (e.g., $25) — emails you when reached

If a bug in your code triggers a million accidental requests, you cap your loss at the hard limit instead of waking up to a $5,000 bill.

Step 4: Create an API Key

1. Go to Dashboard → API Keys
2. Click Create new secret key
3. Name it descriptively (production, dev-laptop, staging)
4. Optional: assign to a Project (limits the key's scope)
5. Click Create
6. Copy the key immediately — it's shown only once

Step 5: Store It Securely

Five non-negotiable rules:

1. Environment variables only. Never hard-code in source files.
2. Server-side only. Never expose to client-side code, even via Next.js public env vars.
3. Separate keys per environment. One for dev, one for staging, one for production.
4. Use a secrets manager. Vercel env vars, AWS Secrets Manager, GCP Secret Manager, or 1Password.
5. Rotate quarterly. Or immediately on suspicion of leak.

Step 6: Your First API Call

Python

pip install openai

# Set OPENAI_API_KEY env var, then:
from openai import OpenAI

client = OpenAI()
response = client.chat.completions.create(
    model="gpt-5",
    messages=[
        {"role": "user", "content": "Hello, world"}
    ]
)
print(response.choices[0].message.content)

Node.js

npm install openai

// Set OPENAI_API_KEY env var, then:
import OpenAI from 'openai'

const client = new OpenAI()
const response = await client.chat.completions.create({
  model: 'gpt-5',
  messages: [{ role: 'user', content: 'Hello, world' }]
})
console.log(response.choices[0].message.content)

2026 Pricing Snapshot

• GPT-5: $5/M input, $15/M output. Flagship model for complex reasoning.
• GPT-4o: $5/M input, $15/M output. Multimodal (image, voice).
• GPT-4o-mini: $0.15/M input, $0.60/M output. Cheap production workhorse.
• GPT-3.5-turbo: $0.50/M input, $1.50/M output. Legacy — most users have moved to 4o-mini.
• o1-preview: $15/M input, $60/M output. Reasoning model for hard problems.

Common Issues

• “You exceeded your current quota”: Add a payment method (free credits aren't enough), or you hit your usage limit.
• “Invalid API key”: Key was revoked or copied with whitespace. Check for trailing newlines.
• Rate limit errors: New accounts have low limits. Pre-pay $5+ to upgrade tiers automatically.
• “Model not found”: Some models require Tier 2+ (after $50 spent). Use GPT-4o-mini until you graduate.

Build AI SaaS With the API Stack That Wins

OpenAI key plus Claude key = the dual-model setup most production AI products run in 2026. Routing requests by task type beats picking one. Our AI SaaS Builder course shows you the exact stack — Claude for code/agents, GPT-4o for multimodal, n8n for orchestration — and the playbook to ship at $10K MRR.

Related Reads

• How to get a Claude API key
• Claude coding complete guide
• Claude API pricing tutorial
• Claude vs GPT-4 comparison
• Claude vs Gemini comparison