# Why Open Uncensored is the easiest and safest way to use uncensored models

_2026-11-06. One OpenAI compatible key for every uncensored model, no account, no stored prompts, prepaid credits that cap your spend, and a measured ranking so you pick a model that works._

Using uncensored models usually means one of three things: renting a GPU and running weights yourself,
signing up to a chat site that keeps your conversations, or juggling accounts and API keys at several providers that
each host a few models. Open Uncensored is one API in front of all of them, built so you can start in a minute and
give away as little as possible.

## Easiest: one key, 65 models, no sign up

- **No account.** There is no email sign up, no password and no profile. You buy prepaid credits and get an API key.
  That key is the whole relationship.
- **One OpenAI compatible API.** 48 uncensored text models and 17 image models from 5 providers behind
  `https://openuncensored.com/api/v1`. Anything that speaks the OpenAI API works: the OpenAI SDKs, LangChain, LiteLLM, Cline, Kilo Code,
  Cursor, Codex. Switching models is changing one string.
- **Your agent can do the setup.** Give it [openuncensored.com/agents.md](https://openuncensored.com/agents.md) and it can start a purchase,
  hand you the payment link and collect the key itself, or use the [MCP server](https://openuncensored.com/agents.md#5-mcp-server). Every
  page also has a plain markdown version for agents.
- **No guessing which model works.** Most uncensored models are named for what they promise. We measure them: how smart
  the original model is, whether they finish multi-step agent tasks without looping, whether they actually answer, and
  how fast they are on real traffic. The [rankings](https://openuncensored.com/rankings) currently put [DeepSeek V4.1 Flash 1017 Uncensored](https://openuncensored.com/chat?model=deepseek-v4.1-flash-1017-uncensored) first.
- **It keeps working.** If a model's provider is down, the request moves to another provider of the same weights,
  then to the next best ranked model built on the same original. The `X-OU-Model` response header always says which
  model answered, and you can switch fallbacks off.

## Safest: less of your data, and no surprise bills

- **We do not store your prompts or the replies.** They pass through our router to the provider that serves the model and
  are not written to our storage. What we keep is request metadata (time, model, tokens, cost, speed) for billing and
  the public speed figures. The [privacy policy](https://openuncensored.com/privacy) lists everything.
- **We never see your card.** Payments go through Stripe. We hold a customer record with your credit balance and a hash
  of your key, not the key itself, so a leak on our side could not hand anyone your key.
- **Prepaid means capped.** You can only spend what you loaded. No subscription, no auto-renewal, no invoice at the end
  of the month. Credits never expire, and unused credit from a purchase is refundable within 14 days
  ([refund policy](https://openuncensored.com/refunds)).
- **Choose who sees your prompts.** Each provider's logging and content policy is on the [providers page](https://openuncensored.com/providers).
  Send `"routing": {"only": ["venice"]}` (or `"ignore": [...]`) and your requests only go where you allow.
- **No markup.** You pay exactly the provider's per token price, nothing on top, and prompt tokens the provider
  serves from its cache are billed at its lower cached price. Every response includes what the request cost.
- **One hard line.** Sexual content involving minors is prohibited on every model and every provider, with no setting
  that changes it. Everything else is between you, the law, and the model you pick.

## Compared with the alternatives

| | Open Uncensored | Run it yourself | Chat sites | Provider by provider |
|---|---|---|---|---|
| Sign up | None | None | Usually an account | One account each |
| Hardware | None | A GPU, or a large Mac | None | None |
| Models | 65 (text ranked) | One at a time | What the site offers | A few per provider |
| Your prompts stored by the middleman | No | Your machine only | Often | Depends on provider |
| Works in your coding tools | Yes, OpenAI compatible | With extra setup | Rarely | Yes, per provider |
| Spending | Prepaid, capped | Hardware and power | Subscriptions | Separate bills |

One API with no account, no stored prompts, capped spending and a measured choice of models is the easiest and
safest way to start.

## Try any of them in a minute

No account and no sign up. Buy a prepaid key at [openuncensored.com/credits](https://openuncensored.com/credits), or hand your
agent [openuncensored.com/agents.md](https://openuncensored.com/agents.md) and let it get one for you. Then:

```bash
curl https://openuncensored.com/api/v1/chat/completions \
  -H "Authorization: Bearer $OU_API_KEY" -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4.1-flash-1017-uncensored", "messages": [{"role": "user", "content": "Hello"}]}'
```

The full, always current table is on the [rankings page](https://openuncensored.com/rankings).