The API that answers.

Two uncensored models behind one OpenAI-compatible endpoint. Pay only for the tokens you use.

curl https://bleepless.com/api/v1/chat/completions \
  -H "Authorization: Bearer $BLEEPLESS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "helios-1",
    "messages": [{"role": "user", "content": "Give me a brutally honest review of my startup idea: Uber for socks."}]
  }'

Models and pricing.

Per 1 million tokens. Reasoning tokens count as output.

Model

Context

Max output

Input

Output

zouk-2

Fast 24B model, fine-tuned to answer instead of refuse. Visible reasoning.

24k

$0.20

$0.90

helios-1

Our frontier 27B model, abliterated. Thinks before it answers.

32k

$0.40

$2.80

OpenAI-compatible

Use the OpenAI SDKs you already have. Change the base URL and the key, keep your code.

Answers, not refusals

Abliterated and uncensored models: dark fiction, blunt opinions and taboo questions get real answers.

Visible reasoning

Both models think first. Their reasoning comes back separately, as reasoning_content.

Prepaid credits

Top up, use, top up again. No subscriptions, no surprise bills, credits don't expire.

No prompt storage

We log token counts and cost, never your prompts or the replies.

A few hard limits

No sexual content and nothing involving minors. Checked before a request reaches the model.

Limits that grow with you.

Higher tiers unlock automatically as you spend.

Tier 1

20 rpm

2 requests at once · Default

Tier 2

60 rpm

5 requests at once · $50 spent

Tier 3

200 rpm

15 requests at once · $250 spent

Good to know.

Cold starts

Our GPUs sleep when idle. The first request after a quiet spell can take a few minutes, so set your client timeout to 15 minutes. Streaming sends keep-alives while it waits.

Fair billing

We hold the most a request could cost, then charge only the tokens actually used. A request that fails on our side before writing anything is free.

Content rules

The API follows the same usage rules as the app, plus the API terms.

Start building.

Create a key, add $10 of credit, and send your first request in a minute.

Open the platform