Uncensored AI API or Model — One Key. Every Frontier Model.

Uncensored AI access for developers: choose an uncensored AI API or an uncensored AI model and run it through one OpenAI- and Anthropic-compatible endpoint — Claude, GPT, Gemini, Grok, DeepSeek, GLM and Qwen, with uncensored and abliterated variants where available.

Get an API key → · Quickstart · Docs · Telegram

An uncensored AI API is a single programmatic endpoint that gives your application access to model variants with fewer built-in refusals than standard hosted models. If you are searching for uncensored AI, an uncensored AI model or an uncensored AI API, the same gateway covers all three: the API is the connection, the model is the engine, and uncensored AI is the category. Instead of maintaining separate SDKs, keys and billing accounts for every provider, you send one request and choose the model family you need.

RouterHub is an OpenAI- and Anthropic-compatible gateway for exactly that workflow. It bundles frontier and open-weight model families — plus uncensored and abliterated variants where available — behind one API key, one balance and one billing model: pay per token, no subscription.

You may also see the same idea called an uncensored LLM API, an unfiltered AI API, a no-refusal AI API, or an abliterated model API. The labels overlap; the practical questions are the same: model coverage, compatibility, billing and behavior.

Uncensored does not mean lawless. Model behavior still depends on the model, version, prompt and provider policy. You stay responsible for lawful, ethical use. See Safety & responsible use.

An uncensored AI model is a model variant trained, fine-tuned or weight-modified to refuse less often than a standard release. The model is the engine; the API is how your application reaches it, and a gateway gives you one key across families.

Uncensored models still vary by version, prompt and provider policy. Evaluate the exact model before deploying. See Uncensored vs. abliterated models.

Model IDs, context limits and live availability are shown in the RouterHub dashboard. Availability changes as providers release new versions — always confirm the current model ID before deploying.

Create an account at routerhub.si. Your API key starts with sk-; the exact Base URL is shown on the dashboard overview.

export ROUTERHUB_API_KEY="sk-..."
export ROUTERHUB_BASE_URL="https://.../v1"   # copy from the dashboard

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["ROUTERHUB_API_KEY"],
    base_url=os.environ["ROUTERHUB_BASE_URL"],
)

response = client.chat.completions.create(
    model="MODEL_ID",  # choose any model ID from the dashboard
    messages=[
        {"role": "system", "content": "You are a precise, helpful assistant."},
        {"role": "user", "content": "Explain the trade-offs of model routing in two paragraphs."},
    ],
)

print(response.choices[0].message.content)

import os
from anthropic import Anthropic

client = Anthropic(
    api_key=os.environ["ROUTERHUB_API_KEY"],
    base_url=os.environ["ROUTERHUB_BASE_URL"],  # use the Anthropic base URL from the dashboard
)

message = client.messages.create(
    model="MODEL_ID",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Write a concise release note for a developer tool."}],
)

print(message.content[0].text)

curl "$ROUTERHUB_BASE_URL/chat/completions" \
  -H "Authorization: Bearer $ROUTERHUB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "messages": [{"role": "user", "content": "Hello from the RouterHub quickstart."}]
  }'

Full setup guides: OpenAI-compatible · Anthropic-compatible · cURL.

Step-by-step guides: Claude Code · Cursor · OpenCode · Codex CLI.

• Pay per token — you are charged for the tokens you actually use.
• No subscription — no monthly minimum and no seat fees.
• Prompt caching — repeated prefixes are billed at the lower cached-input rate.
• Crypto top-ups — add prepaid credits with card or crypto.
• API keys — create separate keys to isolate projects and usage.
• One dashboard — keys, billing, usage and support in one place.

Live model prices are listed in the dashboard and can change when providers update their rates. See Pricing & limits.

• AI safety and red-team research — compare how model versions handle instruction conflicts and policy boundaries.
• Creative writing and worldbuilding — work with fewer mid-scene refusals while keeping your own review process.
• Agentic coding — route Claude, GPT, DeepSeek and GLM through one key in Claude Code, Cursor, Codex or OpenCode.
• Multilingual assistants — test the same prompt across model families without rebuilding your integration.
• Privacy-conscious prototyping — keep billing and keys centralized while you evaluate providers.
• Developer tooling — API keys and one balance make it easier to meter usage per project or client.
• AI-assisted editing and tailoring — rewrite, tighten, adapt tone and tailor copy for different audiences.
• Summarization and document analysis — long-context workflows across model families.
• Chatbot and assistant backends — one API for multiple model families.

• No provider can guarantee that a model will never refuse. Uncensored and abliterated variants reduce refusal behavior; they do not remove the need for evaluation, guardrails or human review.
• Model behavior is version-dependent. A prompt that works on one release may behave differently after an upstream update.
• Provider terms still apply. You are responsible for complying with the terms of the model providers and the laws that apply to you.
• Availability changes. Model IDs and uncensored variants rotate as providers ship updates; check the dashboard before relying on a specific ID.

What is an uncensored AI API?

An uncensored AI API gives developers programmatic access to model variants with fewer refusal behaviors than standard hosted models. RouterHub exposes those variants next to mainstream model families through one OpenAI- and Anthropic-compatible endpoint. Read the full guide →

What is an uncensored AI model?

An uncensored AI model is a model variant with reduced refusal behavior. The model is the engine your app calls; the API is the connection. Read the full explanation.

Is RouterHub an uncensored AI model provider?

RouterHub is an API gateway, not the owner of the models. It provides one key for Claude, GPT, Gemini, Grok, DeepSeek, GLM and Qwen, including uncensored and abliterated variants where available. The model behavior and terms belong to the underlying providers.

Is RouterHub OpenAI-compatible?

Yes. Any client that accepts a custom OpenAI base_url can use RouterHub. The Anthropic-compatible endpoint covers Claude Code and other Anthropic-style clients.

Do I need a subscription?

No. RouterHub is pay-per-token. You top up a balance and spend it across models; there is no monthly subscription.

Can I pay with crypto?

Yes — card and crypto are supported for top-ups.

Are all models uncensored?

No. Standard frontier models are available alongside select uncensored and abliterated variants. Check the model list in the dashboard for current availability.

Can I use RouterHub with Claude Code or Cursor?

Yes. Claude Code uses the Anthropic-compatible endpoint; Cursor uses an OpenAI-compatible provider. Setup steps are in the client guide.

Will an uncensored model never refuse?

No. Refusal behavior varies by model, version, prompt and provider policy. Treat uncensored variants as research and creative tools, not as a guarantee.

Where can I compare model families?

Start with the model catalog and the alternatives guide, then check live availability in the dashboard.

• RouterHub — one key, every frontier model.
• AI Jailbreaks — 35 paste-ready system prompts and agent specs for research.
• Telegram channel — model updates, pricing changes and support.