The project's README from github.com/martina9516/uncensored-ai-models. As one markdown file: README.md.
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.
RouterHub is live: the uncensored model API, live model catalogue and pricing are available at routerhub.si.
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 AI model | The model itself, with reduced refusal behavior. |
| Uncensored AI API | The endpoint that lets your app call that model. |
| API gateway | One key, one bill and one compatible interface across model families. |
Uncensored models still vary by version, prompt and provider policy. Evaluate the exact model before deploying. See Uncensored vs. abliterated models.
| 🔑 One key, every family | Claude, GPT, Gemini, Grok, DeepSeek, GLM and Qwen from a single account. |
| 🔌 Drop-in compatibility | OpenAI SDK, Anthropic SDK, Claude Code, Cursor, Codex CLI, OpenCode and any compatible client. |
| 🧠 Uncensored options | Select uncensored and abliterated variants alongside the standard models. |
| 💳 Pay per token | No subscription. Prompt caching lowers repeated input; card and crypto accepted. |
| 🛡️ Honest limits | Reduced refusals, not a guarantee. Evaluate the model before you deploy. |
| Family | What people use it for | Uncensored variants |
|---|---|---|
| Claude | Long-form reasoning, writing, coding agents | Select variants |
| GPT | General assistants, tools, structured output | Select variants |
| Gemini | Multimodal workflows, research, long context | Select variants |
| Grok | Real-time style answers, ideation | Select variants |
| DeepSeek | Coding, reasoning, cost-efficient agents | Select variants |
| GLM | Agents, coding, bilingual work | Select variants |
| Qwen | Long-document reading, agent workflows | Select variants |
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.
| Client / SDK | Setup |
|---|---|
| OpenAI SDK | Set base_url to the RouterHub base URL |
| Anthropic SDK | Set base_url to the Anthropic-compatible endpoint |
| Claude Code | Point the Anthropic base URL at RouterHub |
| Cursor | Add a custom OpenAI-compatible provider |
| Codex CLI | Use an OpenAI-compatible provider profile |
| OpenCode | Configure a custom provider and model ID |
| LangChain / LlamaIndex | Use the OpenAI-compatible adapter |
| Any OpenAI-compatible app | Change two values: base URL and API key |
Step-by-step guides: Claude Code · Cursor · OpenCode · Codex CLI.
Live model prices are listed in the dashboard and can change when providers update their rates. See Pricing & limits.
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.
| Guide | What it covers |
|---|---|
| What is an uncensored AI API? | Definitions, architecture, who it is for |
| Uncensored vs abliterated models | How the terms differ and what to expect |
| Model catalog | Claude, GPT, Gemini, Grok, DeepSeek, GLM, Qwen |
| Quickstart | Key, base URL, first request in minutes |
| OpenAI-compatible API | SDK, streaming, tools, structured output |
| Anthropic-compatible API | Messages API, Claude Code, system prompts |
| Claude Code, Cursor, OpenCode, Codex | Client-by-client setup |
| Pricing & limits | Pay-per-token, caching, crypto, API keys |
| Use cases | Research, writing, agents, multilingual apps |
| Safety & responsible use | Compliance, evaluation, guardrails |
| FAQ | Long-form answers to common questions |
| Glossary | API, LLM, abliteration, token, gateway |
| Troubleshooting | Auth, model IDs, streaming, rate limits |
| API reference | Request shapes and compatibility notes |
| Alternatives | How to compare AI API providers |
All API access, model IDs, pricing and account setup live on RouterHub:
Open website · Browse models · Get an API key
Star this repository if it helped you evaluate uncensored AI APIs.
RouterHub is an independent API gateway. Model names and trademarks belong to their respective owners. Not affiliated with GitHub, OpenAI, Anthropic, Google, xAI, DeepSeek, Z.ai or Qwen. Use responsibly and follow the terms that apply to you.
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.