uncensored-ai-api

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 Kimi, with uncensored and abliterated variants where available.

Get an API key → · Quickstart · Docs · Telegram bot

---

What is an uncensored AI API?

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.

7API 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.

What is an uncensored AI model?

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.

✨ What you get

🔑 One key, every family	Claude, GPT, Gemini, Grok, DeepSeek, GLM and Kimi 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; BTC, ETH, USDT and LTC accepted.
🛡️ Honest limits	Reduced refusals, not a guarantee. Evaluate the model before you deploy.

🧠 Model families

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
Kimi	Long-document reading, agent workflows	Select variants

Model IDs, context limits and live availability are shown in the 7API dashboard. Availability changes as providers release new versions — always confirm the current model ID before deploying.

Quickstart

1. Get your key and base URL

Create an account at 7api.st. Your API key starts with sk-; the exact Base URL is shown on the dashboard overview.

export SEVEN_API_KEY="sk-..."
export SEVEN_API_BASE_URL="https://.../v1"   # copy from the dashboard

2. OpenAI-compatible request

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["SEVEN_API_KEY"],
    base_url=os.environ["SEVEN_API_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)

3. Anthropic-compatible request

import os
from anthropic import Anthropic

client = Anthropic(
    api_key=os.environ["SEVEN_API_KEY"],
    base_url=os.environ["SEVEN_API_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)

4. cURL

curl "$SEVEN_API_BASE_URL/chat/completions" \
  -H "Authorization: Bearer $SEVEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "messages": [{"role": "user", "content": "Hello from the 7API quickstart."}]
  }'

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

Works with your stack

Client / SDK	Setup
OpenAI SDK	Set base_url to the 7API base URL
Anthropic SDK	Set base_url to the Anthropic-compatible endpoint
Claude Code	Point the Anthropic base URL at 7API
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.

Pricing

• 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 — BTC, ETH, USDT and LTC.
• Subkeys — fund a teammate or client separately from your main balance.
• Referrals — earn 15% of every deposit your referrals make.

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

Use cases

• 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 — subkeys 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.

Honest limits

• 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.

FAQ

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. 7API 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 7API an uncensored AI model provider?

7API is an API gateway, not the owner of the models. It provides one key for Claude, GPT, Gemini, Grok, DeepSeek, GLM and Kimi, including uncensored and abliterated variants where available. The model behavior and terms belong to the underlying providers.

Is 7API OpenAI-compatible?

Yes. Any client that accepts a custom OpenAI base_url can use 7API. The Anthropic-compatible endpoint covers Claude Code and other Anthropic-style clients.

Do I need a subscription?

No. 7API 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 — BTC, ETH, USDT and LTC 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 7API 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.

Documentation

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, Kimi
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, subkeys
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

Related

• 7API — 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.
• Telegram bot — quick access from Telegram.

---

Star this repository if it helped you evaluate uncensored AI APIs.

7API is an independent API gateway. Model names and trademarks belong to their respective owners. Not affiliated with GitHub, OpenAI, Anthropic, Google, xAI, DeepSeek, Zhipu or Moonshot. Use responsibly and follow the terms that apply to you.