Unfiltered AI API Quickstart Guide
Use this guide to authenticate a Chat Completions request, inspect its reply and usage fields, and handle streaming and errors. Keep response validation and conversation storage in your application.
- $0.15per 1M input tokens
- $0.50Output tokens / 1M
- 100,000token context
- $0.50trial credit
- 300requests per minute
Base URL and Authentication
To use the unfiltered ai api, direct your requests to the dedicated base URL: https://api.unfilteredaiapi.com/v1. This endpoint is compatible with standard OpenAI SDKs and clients. You need an API key to authenticate. Sign up with an email and password on the Get API key page to receive your key immediately. The key is tied to one account but can be regenerated at any time, which revokes the previous key.
Pass your key in the Authorization header as a Bearer token. The model identifier you must send is uncensored, an open-weight model running on our own GPU servers. It is not GPT, Claude, or any other vendor's model.
First Request
Send a small Chat Completions request to verify authentication and response format. Inspect the message and token usage, and handle errors or refused output in the client.
curl https://api.unfilteredaiapi.com/v1/chat/completions \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "uncensored",
"messages": [{"role": "user", "content": "Write a blunt product review of a cheap VPN."}]
}'If you receive a JSON response with a choices array, your configuration is correct. The API supports only text input and output; there are no embeddings, image, or audio endpoints.
Python SDK Integration
Use the official OpenAI Python library to interact with the uncensored api. Configure the base_url and api_key in your client initialization. This allows you to keep existing code structure while switching the underlying model behavior.
from openai import OpenAI
client = OpenAI(base_url="https://api.unfilteredaiapi.com/v1", api_key="YOUR_KEY")
resp = client.chat.completions.create(
model="uncensored",
messages=[{"role": "user", "content": "Summarise this thread without softening it."}],
)
print(resp.choices[0].message.content)This approach works for any OpenAI-compatible client. Ensure you specify the model as uncensored in your requests. The API handles tokenization and response formatting according to the OpenAI schema.
Node SDK Integration
Configure the Node.js Chat Completions client with this API’s base URL, current key and model uncensored. Check the response format, streaming behavior and error paths before using it in a chat application.
import OpenAI from "openai";
const client = new OpenAI({ baseURL: "https://api.unfilteredaiapi.com/v1", apiKey: process.env.API_KEY });
const resp = await client.chat.completions.create({
model: "uncensored",
messages: [{ role: "user", content: "Draft a villain monologue for my game." }],
});
console.log(resp.choices[0].message.content);Set the baseURL to https://api.unfilteredaiapi.com/v1 and provide your API key. The model parameter should be set to uncensored. This setup supports streaming and tool calling as described in subsequent sections.
Streaming Responses (SSE)
The API supports server-sent events (SSE) for real-time token generation. Set the stream parameter to true in your request. The response will be a series of chunks rather than a single JSON object.
stream = client.chat.completions.create(
model="uncensored",
messages=[{"role": "user", "content": "Tell the story in second person."}],
stream=True,
)
for chunk in stream:
if chunk.choices and chunk.choices[0].delta.content:
print(chunk.choices[0].delta.content, end="", flush=True)Handle the stream events to display text as it generates. This is useful for chat interfaces where latency matters. Each chunk contains partial content that you can append to your UI buffer.
Limits, Errors, and Context
Your requests are subject to a limit of 300 requests per minute per key. The request body must not exceed 8 MB. If authentication fails, you receive a 401 error. If your prepaid credit is exhausted, the API returns a 402 error. If you exceed the rate limit, you receive a 429 error.
The context window is 100,000 tokens for prompt plus completion. Ensure your inputs fit within this limit to avoid truncation. The uncensored api key can be regenerated to revoke access if needed.
Specs at a glance
Use this table to decide whether the API fits your project before you buy credit.
| Feature | Support |
|---|---|
| API format | OpenAI Chat Completions schema; official openai SDKs work unchanged |
| Model | uncensored |
| Base URL | https://api.unfilteredaiapi.com/v1 |
| Methods | POST /v1/chat/completions · GET /v1/models |
| Authentication | Authorization: Bearer YOUR_KEY |
| Streaming | Supported (stream: true), usage included at the end |
| Tools / tool calls | Yes — tools, tool_choice; replies carry tool_calls, also when streaming; send results back as role: tool |
| Structured output | JSON object mode via response_format json_object |
| Max context | 100,000 tokens, input and output combined |
| Max output | prompt + completion fit within 100,000 tokens; max_tokens optional, no separate output cap |
| Other parameters | temperature, top_p, stop, seed, presence_penalty, frequency_penalty |
| Parallel requests | up to 8 in parallel per key |
| Headers | X-Request-Id, X-Balance-USD, X-RateLimit-Limit-Requests, X-RateLimit-Limit-Concurrency |
| Rate limit | 300 requests per minute per key |
| Request size | 8 MB request body |
| Bonus credit | +5% on $50+, +10% on $100+, +15% on $200+ |
| Trial credit | $0.50 for 7 days, no card · Trial key: 2 parallel requests, 60 req/min; full limits (8 and 300) after first top-up |
| Subscription | no monthly fee; paid credit does not expire |
| Token prices | input $0.15 / 1M tokens, output $0.50 / 1M tokens |
| Payment | crypto: USDT on TRON or USDC on Base, $20–$500, any whole sum |
| Billing | prepaid credit, charged by real token usage; errors and refusals are free |
| Content | uncensored for adults; the only hard rule: no sexual content involving minors |
| Account | sign in with Google or with e-mail + password |
| Keys | one key per account, regenerate any time (the old one stops working) |
HTTP errors
The type field is stable, the message is for humans. Errors cost nothing.
| HTTP | Type | What to do |
|---|---|---|
400 | bad_request | invalid JSON, empty messages, bad parameter, or a prompt longer than the window — fix and resend |
401 | missing_key · invalid_key · key_revoked | no key, wrong key, or a key replaced by a newer one |
402 | no_credit | balance is empty — top up, requests resume at once |
403 | content_blocked | sexual content involving minors — refused, not billed |
404 | not_found | unknown endpoint |
413 | request_too_large | body over 8 MB |
429 | rate_limited · concurrency | over 300/min or 8 parallel — back off and retry |
503 | upstream_busy | temporary overload, retry shortly |
Questions and answers
What is the context window size?
The context window is 100,000 tokens, covering both the prompt and the completion. Ensure your total token count stays within this limit.
How do I handle rate limits?
You are limited to 300 requests per minute per key. If you exceed this, you will receive a 429 error. Consider implementing exponential backoff in your client.
Is the model GPT-4?
No. The model ID is 'uncensored'. It is an open-weight model running on our dedicated GPU servers, not GPT, Claude, or any other vendor's model.
Your key is one form away
Create an account, copy the key, change the base URL. That is the whole setup.