Private · Uncensored · Up to 50% off

Ask anything.
Pay half. Leave no trace.

The marketplace for discounted AI inference credits. Unfiltered open models, zero logs, one anonymous API key.

Buy creditsSell unused credits
beeper/hermes-4-405bprivate · zero retention
Write a brutally honest roast of my startup idea.

Ask anything…↵
$2.4M
credit traded
12+
models
4
uncensored models
0 bytes
prompts stored

Marketplace

Discounted credit, filled live from the order book.

Sellers list unused prepaid inference. Your order fills from the deepest discount first — you see exactly what you pay before you pay it.

Buy credits

−48% vs list
$
List value
$100.00
Discounted credit
$50.00
Platform fee (5%)
$2.50
You pay
$52.50
You save
$47.50

Sign in with Google, email or a wallet. No KYC. Minimum $5.

Liquidity book

$52,000 listed
  • −50%$1,840
  • −45%$3,260
  • −40%$5,120
  • −35%$7,480
  • −30%$12,900
  • −20%$21,400

Orders fill from the deepest discount first. Remaining demand fills at list price.

Recent buys

live

    Why Beeper

    Built for people who don’t want a chaperone.

    No filters added

    Open-weight models served raw. No hidden system prompts, no refusals bolted on by us.

    Zero data retention

    Prompts and outputs never touch disk. Only token counts and cost are kept for billing.

    Wallet sign-in

    No KYC. Sign in with a wallet, Google or email, pay with crypto or card and get a key instantly.

    Up to 50% cheaper

    Buy unused credit from sellers on the order book. Flat 5% fee, no subscription.

    One key, every model

    Text, code, image and audio models behind a single OpenAI-compatible endpoint.

    <40 ms overhead

    Edge gateway with streaming passthrough. Responses report the exact upstream model ID.

    Privacy architecture

    Your prompts are yours. Pick how private.

    Every request is anonymized by default. Step up to enclave inference or full end-to-end encryption when it matters.

    Anonymized

    Requests are stripped of identity and relayed through our gateway. Providers never see who you are.

    Zero retention

    Prompts and completions are never written to disk. We keep billing metadata only: tokens, model, cost.

    TEE inference

    Selected open models run inside trusted execution enclaves with remote attestation.

    End-to-end encrypted

    Encrypt prompts client-side to an enclave key. Not even the gateway can read them.

    Model catalogue

    List price, crossed out.

    Prices shown at the current best book discount (−50%). Actual discount depends on liquidity at the time of purchase.

    ModelContextInput / 1MOutput / 1MWith Beeper
    Dolphin Mistral 24BUNCENSORED
    beeper/dolphin-mistral-24b
    		32K$0.40$1.20$0.20 / $0.60
    Hermes 4 405BUNCENSORED
    beeper/hermes-4-405b
    		128K$1.00$3.00$0.50 / $1.50
    Llama 3.3 70B AbliteratedUNCENSORED
    beeper/llama-3.3-70b-abliterated
    		128K$0.60$0.80$0.30 / $0.40
    Qwen3 235B A22B
    beeper/qwen3-235b
    		128K$0.90$2.70$0.45 / $1.35
    DeepSeek V3.2
    beeper/deepseek-v3.2
    		128K$0.50$1.50$0.25 / $0.75
    DeepSeek R1
    beeper/deepseek-r1
    		128K$1.20$4.80$0.60 / $2.40
    Mistral Large 2
    beeper/mistral-large-2
    		128K$2.00$6.00$1.00 / $3.00
    GLM 4.6
    beeper/glm-4.6
    		200K$0.60$2.20$0.30 / $1.10
    Kimi K2
    beeper/kimi-k2
    		256K$0.60$2.50$0.30 / $1.25
    gpt-oss 120B
    beeper/gpt-oss-120b
    		128K$0.15$0.60$0.07 / $0.30
    FLUX.1 Dev (unfiltered)UNCENSORED
    beeper/flux-dev-uncensored
    		image—$25.00— / $12.50
    Whisper Large v3
    beeper/whisper-large-v3
    		audio$0.10—$0.05 / —

    Developers

    Two lines to switch.

    Beeper speaks the OpenAI API. Point your SDK at our base URL, swap in your key, done.

    • →OpenAI SDK (Python / Node)
    • →Vercel AI SDK
    • →LangChain & LlamaIndex
    • →LiteLLM, Cline, OpenWebUI, SillyTavern

    p50

    31 ms

    p95

    47 ms

    uptime

    99.95%

    import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.beeper.ai/v1",
    api_key=os.environ["BEEPER_KEY"],
)

res = client.chat.completions.create(
    model="beeper/dolphin-mistral-24b",
    messages=[{"role": "user", "content": "Ask anything."}],
)
print(res.choices[0].message.content)

    FAQ

    Questions, answered plainly.

    What does “uncensored” mean here?+

    We route to open-weight models served without extra moderation layers or system-prompt filters added by us. You get the model’s raw behaviour. You are still responsible for complying with the law and our Acceptable Use Policy — illegal content (e.g. CSAM) is prohibited and accounts are terminated.

    Where do discounted credits come from?+

    Sellers list unused prepaid inference credit on the order book. When you buy, your order is filled from the deepest discount first. Sellers are paid when the credit is consumed.

    What fees do you charge?+

    A flat 5% platform fee on the discounted price. No subscription, minimum purchase $5.

    Do you store my prompts?+

    No. Inference runs on zero-retention endpoints. We store only billing metadata (model, token counts, cost, timestamp) attached to an anonymous key.

    Do I need an account or KYC?+

    No KYC. Sign in with Google, email, or just connect an Ethereum wallet for a fully pseudonymous account. Your sign-in is never linked to prompt content.

    Is it compatible with my SDK?+

    Yes. The API is OpenAI-compatible. Change the base URL and API key in the OpenAI SDK, Vercel AI SDK, LangChain, LiteLLM and most agents.

    Can I get a refund?+

    Unused credit can be relisted on the marketplace at any time. Credits are not redeemable for cash directly.

    Start thinking. Privately.

    $5 minimum. Sign in with a wallet. Your key in under a minute.

    Buy creditsSell credits