Open account

Unfiltered Chatbots Like Grok: When an API Fits

Looking for unfiltered chatbots like Grok usually means one of two things: you want a chat app with fewer refusals, or you want an unfiltered model to build with. Apps and APIs solve different problems, and this article helps you pick.

Last updated

Fork in the path from a chat bubble to a finished app box and to open building blocks

Key takeaways

Grok is a chatbot made by xAI, and people name it when they ask for a less restricted assistant. It is a consumer product: you open an app or a site, sign in and talk. Here that means products of that kind, including the many offered by other companies. What they allow, what they cost and what they keep about your chats are set by each company, so check their current terms and pricing.

A hosted text API is a different kind of thing. It has no interface at all. You send a JSON request and receive text. Nothing here is affiliated with xAI, and the service is not the official API of any vendor. It runs one open-weight model, run by an independent provider and tuned to answer without content refusals for lawful adult use.

A fair warning about naming. People use "unfiltered" loosely, to mean anything from fewer content refusals to no limits at all. No hosted service means the second. Here the hard limit is no sexual content involving minors, the target is lawful adult use, and your own product adds its own rules on top. Apps also differ widely in how they apply theirs, so compare what is written in their terms, not what the marketing implies.

Side by side

Consumer chatbotHosted text API (this service)
What you getApp or website with accounts and chatAn endpoint, POST /v1/chat/completions
InterfaceProvidedYou build it, or plug the endpoint into a front end
Persona and rulesSet by the vendor, with some user optionsYours, in the system prompt
MemoryBuilt in, as designed by the vendorYours: you send the history on each call
Images, voiceDepends on the productNone. Text only
Pricing modelCheck the vendor's plansPrepaid, $0.15 input and $0.50 output per 1M tokens
Who handles moderationThe vendorYou

The last row deserves attention. When you use an app, the vendor carries the moderation and age duties. When you build on an API, you do.

There is also a data difference. With a consumer app, your chats live in the vendor's system under the vendor's policy, so read that policy. With this API, prompts are not used for training, and whatever history you keep lives in your own database under your own retention rules. That is more control and more responsibility at once.

Who should pick an app

  • You want to chat, write or roleplay today, with no setup.
  • You want a polished interface, apps on your phone and features such as images or voice, if the product offers them.
  • You do not write code and do not want to.
  • You like the vendor's tone and are fine with their rules.

For these cases, an app is the right answer. Using an API for a one-off chat is more work with nothing gained.

Who should pick an API

  • You are building a bot, a game, a companion app or a writing tool and need to control the persona and the data flow.
  • You want your own rules in the system prompt, not a vendor's defaults.
  • You want costs to follow usage. Pricing here is prepaid, from $20, with no subscription and no expiry.
  • You already use an OpenAI SDK or a client that accepts a custom base URL.

Setup is a few lines. The base URL is https://api.chatbotunfiltered.com/v1, the model ID is uncensored and the key goes in an Authorization: Bearer header.

curl https://api.chatbotunfiltered.com/v1/chat/completions \
  -H "Authorization: Bearer $UF_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "uncensored",
    "messages": [
      {"role": "system", "content": "You are a blunt, witty assistant who answers directly."},
      {"role": "user", "content": "Give me a harsh but fair review of this opening line: It was a dark and stormy night."}
    ]
  }'

The docs cover streaming, tools and JSON mode.

A rule of thumb: if you can describe your project as "I want a bot that does X for Y people", you want an API. If you would describe it as "I want to talk to a bot that does X", you want an app. People often start with the second wish and discover the first once they want to tweak something the app does not allow.

What the API lacks, plainly

It is better to learn the gaps before signing up.

  • No app. There is no chat page, no phone app and no saved chats unless you build them or use a front end that accepts an OpenAI-compatible endpoint.
  • No images, audio or video. Text only. No embeddings and no fine-tuning either.
  • No built-in live data. The model answers from what it was trained on and what you put in the prompt. If you want current information, you supply it, for example through tool calling in your own code.
  • No memory between calls. You manage history, up to the 100,000-token window.
  • Crypto only for payment. USDT on TRC20 or USDC on Base.

Most consumer chatbots are stronger on these points, so if they matter, choose an app.

Using the API inside a chat front end

If you want the API without writing code, there is a middle route. Some chat front ends let you add an OpenAI-compatible endpoint; SillyTavern is one example. The steps are similar in each, but menu names and options differ between versions, so check the front end's current docs.

  1. Open the front end's connection or API settings and choose the OpenAI-compatible, or custom, option.
  2. Enter https://api.chatbotunfiltered.com/v1 as the base URL.
  3. Paste your key and set the model to uncensored.
  4. Set the context size to something sensible. The window is 100,000 tokens for prompt plus reply, but larger prompts cost more on every turn.
  5. Turn on streaming if the front end offers it, and send a test message.

You then get an interface, character cards and saved chats from the front end, and the model from the API. Each part does what it is good at. Costs follow the tokens the front end sends, so watch the first long chat and check the balance, either in the X-Balance-USD header or with GET /v1/balance.

What it costs to try

Each turn is priced by tokens. A turn with a 2,000-token prompt and a 300-token reply costs 2,000 × $0.15 / 1,000,000 + 300 × $0.50 / 1,000,000 = $0.0003 + $0.00015 = $0.00045. New users can claim $0.50 of trial credit with one click on the API key page, which covers about 1,111 such turns ($0.50 ÷ $0.00045). It is valid for 7 days, needs no card and is limited to one per person.

For a deeper look at costs at scale, see the site's NSFW chatbot cost and trade-off analysis, and the pricing page for the top-up table. To start, get a key. Failed or refused requests are not charged, and prompts are not used for training.

One more point on responsibility. The model is tuned for lawful adult use, and the one hard limit is no sexual content involving minors. If you build a public product on it, age checks and moderation are your job.

Compare that with what you pay now. If you use an app on a plan, take the monthly price from its site, divide by $0.00045 and you get the number of turns the same money buys here. That is a fair check of whether the move is worth the setup effort. Remember that the API turns are bare text, while an app plan may include features you would then have to build or give up.

Frequently asked questions

Is this the Grok API?

No. It is not affiliated with xAI. It is a hosted open-weight model run by an independent provider, reached through an OpenAI-compatible API.

Are there unfiltered chatbots like Grok that I can just use?

Yes, many apps exist, and their rules and prices change, so check each one's current terms. If you want to build your own bot instead, an API is the building block.

Does the API generate images?

No. It is text only. There is no image, audio or video generation.

Can I connect it to a chat front end?

If the front end accepts an OpenAI-compatible endpoint, enter the base URL, the key and the model ID uncensored.

Get your API key instantly

Register, copy your key, update your base URL. Setup is complete.