LLM Free API Key: What the Trial Credit Covers

The short answer: one small free allowance, no free tier

If you are searching for an llm free api key, here is the plain version. This service does not offer a permanent free plan with a monthly quota. What it offers is a one-time trial: after you sign up, the API key page has a button that adds $0.50 of credit to your account. You can claim it once per person, it needs no card, and it stays valid for 7 days.

Fifty cents sounds like very little until you convert it into tokens. The model uncensored costs $0.15 per 1M input tokens and $0.50 per 1M output tokens, so $0.50 is enough to send several hundred real requests and read the answers. That is plenty to decide whether the model suits your app, your bot or your story project before you put any money in.

It helps to be clear about what the trial is not. It is not an unlimited key, it is not a shared public key, and it does not renew. When it runs out or expires you either top up or stop. The rest of this page explains how far the credit goes, how to claim it, and what the paid path costs.

What $0.50 of trial credit buys

Cost depends on two numbers per request: how many tokens you send (input) and how many the model writes back (output). Output is more expensive per token, so long answers cost more than long questions. Here is the arithmetic for four common request shapes, all computed from the two prices above.

Request shape	Input / output tokens	Cost per request	Requests per $0.50
Short chat turn	500 / 200	$0.000175	about 2,850
Typical chat with history	2,000 / 500	$0.00055	about 900
Long document summary	20,000 / 1,000	$0.0035	about 140
Full 100,000-token window	100,000 / 2,000	$0.016	about 31

Take the typical row as an example. 2,000 input tokens at $0.15 per 1M is 2,000 / 1,000,000 x $0.15 = $0.0003. 500 output tokens at $0.50 per 1M is 500 / 1,000,000 x $0.50 = $0.00025. The sum is $0.00055 per request, and $0.50 divided by $0.00055 is about 909 requests.

The last row is the extreme case. The context window is 100,000 tokens, prompt plus completion, and filling it with 100,000 input tokens costs $0.015 on its own. Few apps do that on every call, but it is useful to know that even a maxed-out prompt costs under two cents.

How to claim the trial credit

1. Open the API key page and sign up with Google, or with an email and password.
2. After sign-up you land on the key page. Create your key and copy it somewhere safe.
3. Press the trial button on that page. The $0.50 appears as your balance straight away.
4. Call the API within 7 days. Unused trial credit expires after that.

No card is requested at any step. The trial is limited to one per person, so a second account will not get a second grant. Every response includes an X-Balance-USD header, which is the easiest way to watch the balance fall as you test.

A first test call looks like this. It uses the standard OpenAI chat completions format, so any OpenAI client works once you point it at the base URL below.

Copycurl https://api.getaiapikey.com/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "uncensored",
    "messages": [{"role": "user", "content": "Write a two-line story opening."}],
    "max_tokens": 100
  }'

The quickstart in the docs has the same call in Python and Node, plus streaming and tool calling.

Limits that apply to a trial key

Trial keys run with tighter rate limits than full keys. This matters if you were planning to load-test during the trial.

Key type	Concurrent requests	Requests per minute
Trial	2	60
Full (after a top-up)	8	300

If you go over, the API answers with HTTP 429. The error code is rate_limited when you exceed the per-minute count and concurrency when too many requests are in flight at once. A streamed request holds its concurrency slot until the stream finishes, so two long streams at the same time can fill a trial key.

For testing, 2 concurrent requests and 60 per minute are enough to try prompts by hand, run a small script, or point a single chat front end at the endpoint. They are not enough to benchmark a busy app, and the trial is not meant for that.

What happens when the trial is used up

From here the model is pay as you go. You top up prepaid credit, and each request deducts the exact token cost from your balance. There is no subscription and no monthly minimum. Failed or refused requests are not charged, and credit does not expire.

Top-ups are any whole amount from $20 to $500. Larger top-ups carry a bonus on the credit you receive:

Top-up	Bonus	Credit added
$20	none	$20.00
$50	+5%	$52.50
$100	+10%	$110.00
$200	+15%	$230.00
$500	+15%	$575.00

You can pay by card (Visa, Mastercard, Apple Pay, Google Pay) or with crypto (USDT on TRC20, USDC on Base). Full details sit on the pricing page. A full key has limits of 8 concurrent requests and 300 per minute.

What $20 covers in practice

The smallest top-up is $20, and it goes a long way at these rates. Using the same four request shapes:

• Short chat turns (500 in, 200 out): $20 / $0.000175 is about 114,000 requests.
• Typical chat with history (2,000 in, 500 out): $20 / $0.00055 is about 36,000 requests.
• Long document summaries (20,000 in, 1,000 out): $20 / $0.0035 is about 5,700 requests.
• Full-window calls (100,000 in, 2,000 out): $20 / $0.016 is 1,250 requests.

Another way to see it: $20 buys about 133 million input tokens if you only count input, or 40 million output tokens if you only count output.

A small bot answering 300 messages a day at the typical shape costs 300 x $0.00055 = $0.165 per day, or about $5 over 30 days. Your own numbers will differ, since history length drives the input side. The AI API checklist on this site lists what else to compare when you pick a provider.

Be careful with other free key offers

Search results for a free LLM key include lists of keys, shared accounts and generators. Treat these with suspicion. A key posted publicly is usually revoked quickly, or it belongs to someone who will see your prompts in their logs. Pasting private text through a stranger's key is the opposite of private.

A safer pattern is a key issued to you alone, with a published price and a trial you can verify yourself. Here the trial is small on purpose, and the data handling is stated: prompts are not used for training. You can read the privacy page before you sign up.

Also keep the product limits in mind. This is a text API only: no images, audio, video or embeddings, and no fine-tuning. It serves one open-weight model run by an independent provider, tuned to answer without content refusals for lawful adult use. The one hard rule is that sexual content involving minors is never allowed.

Stretching the trial credit

A few habits make $0.50 last longer while you evaluate:

• Set max_tokens during tests so a runaway answer cannot burn credit. It is optional, but useful while you experiment.
• Trim the chat history you resend. Input tokens are cheap, but a growing history multiplies across every turn.
• Use a fixed seed and low temperature when comparing prompts, so differences come from your wording and not from sampling noise.
• Test with short inputs first, then run one or two long ones to check how the model handles a big context.

More guides

• How to Install an LLM in Python: Local or API
• LLM API Key for a Discord Bot: Setup Guide
• AI API Checklist