OpenAI
The most well-known AI provider — GPT-4o, GPT-4o-mini, and more.
OpenAI is the industry-standard AI provider. ContentAI supports GPT-4o, GPT-4o-mini, GPT-4 Turbo, and GPT-3.5 Turbo out of the box. OpenAI does not have a free tier for API use — you'll need to add billing credit to your account.
Get a key
- Go to platform.openai.com/api-keys
- Sign in and complete billing setup ($5 minimum top-up)
- Click Create new secret key
- Name it, scope it if you want, and copy — starts with
sk-orsk-proj-
Key shown once
OpenAI only shows you the full key once at creation. Save it somewhere safe (1Password, Bitwarden) before closing the dialog.
Models available in ContentAI
| Model | Context | Price (approx) | Best for |
|---|---|---|---|
gpt-4o-mini | 128k | $0.15/M in · $0.60/M out | ⭐ Budget default |
gpt-4o | 128k | $2.50/M in · $10/M out | Best overall quality |
gpt-4-turbo | 128k | $10/M in · $30/M out | Older flagship, long tasks |
gpt-3.5-turbo | 16k | $0.50/M in · $1.50/M out | Very cheap, short outputs |
Prices change — check openai.com/pricing for current rates.
Quality notes
- GPT-4o is the sweet spot for professional content. Consistent, follows instructions, handles nuance.
- GPT-4o-mini is shockingly good for its price. Most ContentAI users can stick with it 80% of the time.
- GPT-4 Turbo has a reputation for being slightly more "formal"; some writers prefer it.
- GPT-3.5 Turbo is legacy — only use it for very short, cheap outputs.
Setup in ContentAI
- Settings → OpenAI card → paste key
- Pick gpt-4o-mini to start
- Click Test
- (Optional) Set as active provider
Monitoring usage
OpenAI has a live usage dashboard at platform.openai.com/usage. Set a monthly budget and alerts there to avoid surprise bills.
Project vs. user keys
OpenAI offers two key types:
- Project keys (
sk-proj-...) — scoped to a specific project; recommended - User keys (
sk-...) — account-wide; fine for personal use
Both work with ContentAI.
Common errors
| Error | Fix |
|---|---|
Incorrect API key provided | Key is wrong, revoked, or from a different account |
You exceeded your current quota | Add billing credit |
The model ... does not exist or you do not have access | Upgrade tier or pick a different model |
Rate limit reached | Wait a moment; move to a higher usage tier if frequent |
Using your OpenAI key server-side
Same pattern as Groq — set OPENAI_API_KEY in .env.local and fall back to it when no client key is provided. Combine with auth + rate limiting before exposing a shared key publicly.
Next: Anthropic →