Every frontier model,
one magic key.
GPT, Grok and more answer to the same OpenAI-compatible endpoint. See every price up front, top up when you need, and switch models by changing one string.
From sign-up to first response in minutes
Sign up with a username and password. No subscription and no monthly fee.
Top up from $5 with a card through Stripe Checkout. Credit is shared across every model.
Create a key in the dashboard, set the base URL to magicai.lol/v1 and keep using your OpenAI SDK.
Models you can call today
Prices are the per-token rates the gateway actually charges. Speed figures come from real requests through MagicAI and appear once a model has enough samples.
OpenAI's top publicly available API model, with reasoning support and up to 128K output tokens.
Low-cost member of the GPT-6 family for high-volume workloads.
Mid-tier GPT-5.6 model with configurable reasoning effort and up to 128K output tokens.
Plus 15 more models listed in the catalog as Coming soon — shown for reference at official list prices, not callable yet. Browse all →
Your existing code already works
MagicAI speaks the OpenAI Chat Completions format, including streaming. Change the base URL and the key; that's the whole migration.
curl https://magicai.lol/v1/chat/completions \
-H "Authorization: Bearer $MAGICAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "grok-4.7",
"messages": [{"role": "user", "content": "Explain DNS in one sentence."}]
}'Built to stay out of your way
A single sk- key works for every live model. Rotate or delete keys from the dashboard at any time.
Every model shows its input and output rate per million tokens.
Add credit through Stripe when you need it. No plan to pick, no minimum monthly spend.
Every request is logged with model, token counts and cost, so you can see exactly where credit went.
If an upstream attempt fails, the gateway retries before returning an error to you.
Latency and throughput numbers on model pages are computed from real traffic — with the sample size shown.
What will it cost?
Plug in your workload. The calculator uses the live per-token rates of each model.
Questions, answered
Is MagicAI compatible with the OpenAI SDK?
Which models can I use?
How does billing work?
What happens when I run out of credit?
Do you store my prompts?
Where do the speed numbers come from?
Ready to make your first call?
Create an account, add $5 of credit and you're one request away from every live model.