How to Make an Unrestricted AI With an API
The fastest way to learn how to make an unrestricted AI is to skip training and use a hosted model that already answers without content refusals. You sign up, get a key, and call it like any OpenAI-compatible API.

Updated
Key points
- No GPUs or training needed: you call a hosted open-weight model through an API.
- Base URL https://api.unrestrictedaimodels.com/v1, model
uncensored, key in a Bearer header. - Your persona and rules go in the system prompt; the app logic is yours.
- A free-to-start $0.50 trial credit covers roughly 1,500 short requests.
What you are really building
When people go looking for this, they usually mean one of two things. Some want to train or fine-tune a model so that it stops refusing requests. Others want a chatbot, a writing tool or a roleplay app that does not shut down mid-scene. The second goal is far easier to reach.
Running an open-weight model yourself means renting or owning GPUs, installing serving software and keeping it updated. Fine-tuning adds data work on top. A hosted API removes all of that: an independent provider runs one open-weight model tuned to answer without content refusals for lawful adult use, and you send it requests.
What you build is the application around it: the persona, the memory, the interface. This service does not offer fine-tuning, images, audio or embeddings. It is text only, and sexual content involving minors is not allowed.
Step 1: get a key
- Sign up on the API key page with Google, or with an email and password.
- On the key page, claim the trial credit with one click. It is $0.50, valid for 7 days, one per person, and needs no card.
- Copy the key and keep it out of your source code. Use an environment variable.
Trial keys allow 2 concurrent requests and 60 requests per minute. A full key, which you get after topping up, allows 8 concurrent requests and 300 per minute.
Step 2: send the first request
Start with curl to confirm everything works:
curl https://api.unrestrictedaimodels.com/v1/chat/completions \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "uncensored",
"messages": [
{"role": "user", "content": "Write a two-sentence opening for a noir detective story."}
]
}'
You get back standard JSON with the reply in choices[0].message.content. Response headers include X-Request-Id and X-Balance-USD, which shows your remaining credit.
The same call with the Python SDK (pip install openai):
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.unrestrictedaimodels.com/v1",
api_key=os.environ["API_KEY"],
)
resp = client.chat.completions.create(
model="uncensored",
messages=[
{"role": "user", "content": "Write a two-sentence opening for a noir detective story."}
],
)
print(resp.choices[0].message.content)
In Node, install openai and pass baseURL and apiKey to the constructor in the same way. Any other OpenAI-compatible library works once those two values are changed.
Step 3: give it a personality
The system message is where an AI gets its character. You decide the tone, the rules of the world and the limits of the persona. For an AI that has no preset behavior, this message is the entire definition.
messages = [
{
"role": "system",
"content": (
"You are Vesper, a dry-witted ship computer in a sci-fi serial. "
"Stay in character. Keep replies under 150 words. "
"Address the user as Captain."
),
},
{"role": "user", "content": "Vesper, report on the engine room."},
]
resp = client.chat.completions.create(
model="uncensored",
messages=messages,
temperature=0.9,
top_p=0.95,
)
print(resp.choices[0].message.content)
Temperature, top_p, stop, seed and the repetition penalties are passed through. Raise temperature for looser fiction and lower it for factual or structured tasks. Setting a seed helps when you want to compare two prompts fairly.
Step 4: add memory and streaming
The API remembers nothing between calls. To make a chat, append each user message and each reply to the same list and send the whole list every time. The context window is 100,000 tokens, so long conversations fit, but you pay input price on the history with each request.
For a chat interface, stream the answer so text appears as it is written:
stream = client.chat.completions.create(
model="uncensored",
messages=messages,
stream=True,
)
reply = ""
for chunk in stream:
if not chunk.choices:
continue
piece = chunk.choices[0].delta.content or ""
reply += piece
print(piece, end="", flush=True)
messages.append({"role": "assistant", "content": reply})
If your app needs the model to fetch data or trigger actions, tool calling and JSON mode (response_format set to {"type": "json_object"}) are both supported. See the documentation for the full list of options.
When the first request fails
Most first-run problems come from four mistakes:
- Wrong base URL. It must end in
/v1:https://api.unrestrictedaimodels.com/v1. If you paste the full path to/chat/completionsinto an SDK that adds it itself, the path doubles. - Bad key. The header is
Authorization: Bearer YOUR_KEY. A stray space or newline from copying is enough to break it. - Wrong model name. Use
uncensored. You can confirm what the server offers withGET /v1/models. - Rate limits. A trial key allows 2 concurrent requests, so firing off ten at once from a script will get some turned away. Add a small queue or retry with a pause.
Failed requests are not charged, so debugging does not eat your credit. Keep the X-Request-Id header from a response so you can trace a specific call in your logs.
Costs and top-ups
You pay per token from prepaid credit: $0.15 per 1M input tokens and $0.50 per 1M output tokens. There is no subscription, credit never expires, and failed or refused requests are not charged.
The trial shows what that means in practice. A request with 500 input tokens and 500 output tokens costs 500 x $0.15 / 1M + 500 x $0.50 / 1M = $0.000075 + $0.00025 = $0.000325. Divide $0.50 by that and you get about 1,500 requests.
When you need more, top up any whole amount from $20 to $500 by crypto (USDT on TRC20, USDC on Base). Bonuses add +5% from $50, +10% from $100 and +15% from $200, so $200 becomes $230 of credit. The pricing page has the details.
Before you publish an app
Keep your key on a server, never in browser code. Add your own rate limiting so one user cannot use up your balance. Tell your users that the service is for lawful adult use, and decide how your app handles age checks. Prompts sent to this API are not used for training, but your own logs are your responsibility.
The pay-per-token chatbot page shows how costs add up for a bot with many users.
More guides
Questions and answers
Do I need to train a model to make an unrestricted AI?
No. You can call a hosted open-weight model that is already tuned to answer without content refusals for lawful adult use, and build your app around it.
Which SDK should I use?
Any OpenAI SDK. Set the base URL to https://api.unrestrictedaimodels.com/v1, use your key, and set the model to uncensored.
Is there a free way to try it?
New users can claim $0.50 of trial credit on the API key page after sign-up. It is valid for 7 days, one per person, and needs no card.
Does it generate images or voice?
No. The service is text only. It has no image, audio or video generation, embeddings or fine-tuning.
Your key is one form away
Create an account, copy the key, change the base URL. That is the whole setup.
Get API key