Free AI API Key: What to Expect
Searching for a free AI API key often leads to leaked keys or traps. There is a clearer path: a $0.50 free trial credit claimed by yourself, no card needed.

Updated

Key points
$0.50 free trial credit, one click on the API key page, no card needed, valid for 7 days.
One trial per person, with stricter limits: 2 parallel requests, 60 per minute.
With $0.50 you can make about 900 medium-sized chat requests.
Avoid free keys from third parties: they may die, leak, or be used to read your prompts.
What "free" means here
Regarding free AI API keys, the honest answer is this: there are no truly free unlimited keys, but there is a free trial credit sufficient to test the service. New users can claim a $0.50 free trial credit with one click on the API key page after signing up. No card is needed, and the credit is valid for 7 days.
The trial is limited to one per person. The goal is simple: you can verify that this endpoint fits your application before topping up your balance. After that, you are free to decide whether to continue or not.
Steps to claim the free trial credit
- Open the API key page and sign up. You can sign in with Google, or with email and password.
- After signing in, find the claim free trial credit button on the same page and click it once.
- Copy your API key and store it securely, for example in a password manager or environment variable.
- Note the date. The free trial credit expires after 7 days, so plan your testing within that week.
One note: the claim button is only available for new accounts that have not yet used their trial quota. If you do not see it, ask via the Support page before creating another account, as the limit is one per person.
What you can do with $0.50
The service price is $0.15 per 1 million input tokens and $0.50 per 1 million output tokens. From this rate we can estimate the value of the free trial credit.
| Request type | Input / output tokens | Cost per request | Estimated number from $0.50 |
|---|---|---|---|
| Short reply | 500 / 200 | $0.000175 | about 2,857 |
| Chat with context | 2,000 / 500 | $0.00055 | about 909 |
The calculation: 0.50 / 0.000175 = 2,857, and 0.50 / 0.00055 = 909. That is enough to test prompts, try streaming, check function calling, and run several hundred test conversations.
Trial keys have lower limits than full keys: 2 parallel requests and 60 requests per minute. If you exceed these, the API returns HTTP 429 with concurrency or rate_limited. Streaming holds one slot until the stream finishes, so two running streams already meet the concurrency quota.
First call
This service is OpenAI-compatible, so the standard OpenAI SDK works after you change the base URL and key. The model ID is uncensored. Start with curl to ensure the key works:
curl https://api.apitanpasensor.com/v1/chat/completions \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "uncensored",
"messages": [{"role": "user", "content": "Halo, perkenalkan dirimu dalam satu kalimat."}]
}'
Python version with the openai SDK, reading the key from an environment variable:
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.apitanpasensor.com/v1",
api_key=os.environ["API_KEY"],
)
resp = client.chat.completions.create(
model="uncensored",
messages=[{"role": "user", "content": "Halo, perkenalkan dirimu dalam satu kalimat."}],
)
print(resp.choices[0].message.content)
And the Node.js version:
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.apitanpasensor.com/v1",
apiKey: process.env.API_KEY,
});
const resp = await client.chat.completions.create({
model: "uncensored",
messages: [{ role: "user", content: "Halo, perkenalkan dirimu dalam satu kalimat." }],
});
console.log(resp.choices[0].message.content);
To check your remaining credit, look at the X-Balance-USD header in the response, or call GET /v1/balance. The full parameter list is in the documentation.
7-day test plan
Limited credit is most useful for clear questions. The following list arranges sensible tests before you decide to top up your balance.
Day 1: ensure connection and quality
Run the first call, then try five to ten prompts that represent your actual work. Compare the answers with what you need, not with generic examples.
Days 2 to 3: features you rely on
Streaming uses SSE and reports token usage at the end. Python example:
from openai import OpenAI
import os
client = OpenAI(base_url="https://api.apitanpasensor.com/v1", api_key=os.environ["API_KEY"])
stream = client.chat.completions.create(
model="uncensored",
messages=[{"role": "user", "content": "Tulis tiga ide nama untuk toko roti."}],
stream=True,
)
for chunk in stream:
if chunk.choices and chunk.choices[0].delta.content:
print(chunk.choices[0].delta.content, end="", flush=True)
If your app needs structured output, try JSON mode with response_format set to {"type": "json_object"}. If you use functions, try tools and tool_choice, then send the results back with the tool role.
Days 4 to 7: load and cost
Measure the average cost per request from usage, then multiply by your estimated traffic. Also test what happens when you hit the trial key limit, so that your retry code is ready before switching to the full key.
Save some credit: one day of light testing uses only a small fraction of $0.50. Do not spend it all on the first day.
JSON mode example
JSON mode is useful if the model output will be processed by a program, such as filling out forms or tables. Also mention the word JSON in the prompt to make the requested format clear.
curl https://api.apitanpasensor.com/v1/chat/completions \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "uncensored",
"response_format": {"type": "json_object"},
"messages": [{"role": "user", "content": "Beri tiga ide nama toko roti dalam JSON dengan kunci nama dan alasan."}]
}'
Notes on errors during testing
If you receive 429, read the error code. concurrency means too many parallel requests are running, rate_limited means too many per minute. Neither indicates a broken key. Wait a short while and retry, and if there is a Retry-After header, follow its value. Failed requests do not reduce your trial credit.
Keep your API key secure
Leaked keys are a real risk, even for small credits. Others can spend your balance, and with trial keys, they can spend the credit in a matter of minutes. The following habits prevent most cases:
- Store the key in environment variables or a
.envfile, and add that file to.gitignorebefore the first commit. - Do not put the key in front-end code, mobile apps, or public repositories. Anyone can read it from there.
- Call the API from your own server, then provide an internal endpoint for app users.
- Do not share screenshots that show the key, including in forums and chat groups.
- If the key has already been seen publicly, assume it is leaked. Stop using it and contact Support for the next steps.
One more habit: measure usage after a few days. If the balance drops faster than calculated from your logs, it is a sign someone is using your key.
Beware of free keys from other sources
In chat groups and public repositories, keys claimed to be free sometimes circulate. You should avoid them for several common reasons:
- The key often comes from someone else's leaked account. The owner can revoke it at any time, and your app stops working without warning.
- The key owner or intermediary can read every prompt you send. Do not send personal data, secret keys, or client documents through such channels.
- You have no certainty about limits, availability, or where to complain when issues arise.
- Using access that is not yours may violate the original provider's rules.
In contrast, on this service prompts are not used to train the model. The full policy is on the privacy page.
After the free trial credit runs out
When $0.50 or 7 days run out, you have two options: stop, or top up. Top-up starts at $20 and can go up to $500, with a +5% bonus from $50, +10% from $100, and +15% from $200. Payment via crypto: USDT on TRC20 or USDC on Base. Top-up credit never expires.
The paid key is also more relaxed: 8 parallel requests and 300 requests per minute. Rate details are on the pricing page.
Other guides
Frequently asked questions
Does test credit require a payment method?
No. The $0.50 free trial credit is claimed with one click on the API key page after signing up, without a card and without any payment.
How long is the free trial credit valid?
7 days from claiming. After that, remaining free trial credit cannot be used.
Can I claim the free trial more than once?
No. The trial is limited to one per person.
Can this API be used with the OpenAI SDK?
Yes. Just change the base URL to https://api.apitanpasensor.com/v1, using your key, and set the model to uncensored.
Can this service generate images?
No. This API is for text only, with no image, audio, video, embedding, or fine-tuning.
Your key needs just one step
Create an account, copy the key, change the base URL. That is the entire setup process.