Skip to content

The LLM API gateway for Vietnamese developers

One API key, many top models.

Any puts many LLMs behind one OpenAI-compatible endpoint — just change base_url. Top up by VND bank transfer via VietQR; your wallet and prices are in USD, billed by the tokens you actually use.

  • OpenAI SDK compatible
  • Automatic VietQR top-ups
  • No monthly fees
Change base_url and you're done
from openai import OpenAI client = OpenAI(    base_url="https://<host>/v1",    api_key="sk-kompa-...",)resp = client.chat.completions.create(    model="deepseek-ai-deepseek-v4-1-flash",    messages=[{"role": "user", "content": "Hello!"}],)

The highlighted line is the only change from your existing OpenAI code.

Features

Everything you need to bring LLMs into your product

A lightweight API gateway: you focus on building your app, Any takes care of provider connections, billing and usage tracking.

  • OpenAI compatible

    Use the OpenAI SDK (Python, JS) as is, or any library that supports the OpenAI API. Only the base_url and API key change.

  • Many models, one endpoint

    Chat and embedding models from multiple providers behind a single address. Pick one with the model parameter — no need for multiple accounts.

  • Streaming

    Supports stream: true to receive tokens in real time, ideal for chatbots and interactive apps.

  • Pay for actual tokens

    No monthly fees, no fixed plans. You're only charged for the input and output tokens you actually use.

  • Automatic VietQR top-ups

    Scan the VietQR code with your banking app; your balance is credited automatically once the transfer is reconciled.

  • API keys with limits

    Create multiple keys and set a spending limit, a requests-per-minute limit and a list of allowed models for each one.

  • Detailed usage dashboard

    Track tokens and spend by model, by key and by day, with a transparent wallet transaction history.

  • Affiliate & reseller

    A partner program for anyone who refers customers or resells the service. Sign up to join.

Get started

Up and running in 4 steps

From sign-up to your first request in just a few minutes.

  1. Step 1: Create an account

    Sign up with your email in seconds. No credit card required.

  2. Step 2: Top up with VietQR

    Choose an amount and scan the code with your banking app. Your wallet is credited automatically once the transfer is matched.

  3. Step 3: Create an API key

    Give it a name and, if needed, a spending limit and a requests-per-minute limit. The key is shown only once.

  4. Step 4: Call the API

    Point the OpenAI SDK at Any's base URL, pick a model and start making requests.

Models

Models available now

List prices in USD per 1M tokens, updated live from the platform.

  • deepseek-ai/DeepSeek-V4.1-Flash

    deepseek-ai-deepseek-v4-1-flash
    Chat
    Input / 1M
    $0.271318
    Output / 1M
    $0.697675
    Context window
    Context — tokens
  • deepseek-ai/DeepSeek-V4-Flash

    deepseek-ai-deepseek-v4-flash
    Chat
    Input / 1M
    $0.11628
    Output / 1M
    $0.232559
    Context window
    Context — tokens
  • google/gemma-3-12b-it

    google-gemma-3-12b-it
    Chat
    Input / 1M
    $0.0969
    Output / 1M
    $0.193799
    Context window
    Context — tokens
  • google/gemma-3-27b-it

    google-gemma-3-27b-it
    Chat
    Input / 1M
    $0.07752
    Output / 1M
    $0.174419
    Context window
    Context — tokens
  • google/gemma-3-4b-it

    google-gemma-3-4b-it
    Chat
    Input / 1M
    $0.05814
    Output / 1M
    $0.124032
    Context window
    Context — tokens
  • ibm-granite/granite-4.2-3b

    ibm-granite-granite-4-2-3b
    Chat
    Input / 1M
    $0.03876
    Output / 1M
    $0.143411
    Context window
    Context — tokens

Billing

Transparent pricing in USD

You only pay for the tokens you use. No hidden fees and no international card needed: top up by VND bank transfer.

  • Priced per 1M tokens

    Every model has a USD list price per 1M input tokens and per 1M output tokens, billed down to the individual token.

  • Input and output billed separately

    The tokens you send (prompt) and the tokens the model returns (completion) have different unit prices, clearly shown in the pricing table.

  • Hold first, then charge

    Each request places a hold for its estimated maximum cost. When it finishes, only the actual tokens are charged and the rest returns to your available balance immediately.

  • Prepaid, no subscription

    Spend only what you top up. No monthly fees, no minimum commitment. Every wallet transaction is recorded.

Worked example

Hypothetical prices, not those of any specific model.

Input price
$0.40 / 1M
Output price
$1.60 / 1M
2,000 input tokens
$0.0008
1,000 output tokens
$0.0016
Request cost
$0.0024

Formula: tokens × unit price ÷ 1,000,000, rounded up to $0.000001. Plan prices may differ from list prices.

See model pricing

Partners

Affiliate & reseller program

Refer customers with your own referral code, or buy at wholesale prices to resell the service to your customers. Commissions and wholesale prices are agreed with each partner.

  • Affiliate

    Earn commission on the usage of customers you refer, tracked in the partner portal.

  • Reseller

    Your own wallet at wholesale prices — package and resell to your own customers.

  • Kompa Group

    Any — a Kompa Group product. A company registered in Vietnam.

FAQ

Frequently asked questions

What is Any?

Any is an API gateway for large language models (LLMs): one OpenAI-compatible endpoint to call many models from many providers, prepaid, with top-ups by bank transfer via VietQR. The product is operated by Kompa Group.

Will my existing OpenAI SDK code work?

Yes. Point base_url at Any, use your Any API key and pick a model from the pricing table. Chat completions (including streaming) and embeddings use exactly the OpenAI request/response format.

How am I billed?

By the actual tokens of each request: input tokens × input price + output tokens × output price, with unit prices in USD per 1M tokens. There are no monthly or subscription fees.

How do I top up?

Under Top up, enter an amount in USD or VND, then scan the VietQR code with your banking app (a VND bank transfer, no international card needed). Your wallet balance is in USD, converted at the USD/VND rate locked when the QR code is created; once the transfer is reconciled, your wallet is credited automatically.

What does "hold first, then charge" mean?

Before calling the model, the system places a hold for the request's estimated maximum cost. When the request completes, only the actual tokens are charged and the rest of the hold is returned to your available balance.

What happens if my balance runs out?

New requests are rejected with error code 402 (in the OpenAI error format) until you top up. You are never charged more than the balance you've added.

Is the price I pay the same as the public pricing?

The public pricing shows list prices. Some accounts on a custom pricing plan may have different unit prices; after you log in, the Models page in the dashboard shows the exact prices that apply to you.

Can I cap spending per API key?

Yes. Each key can have a spending limit, a requests-per-minute limit and a list of allowed models. You can revoke a key at any time.

How do I join the affiliate or reseller program?

Sign up for an account or contact ai@kompa.ai and we'll be in touch. Commission rates and wholesale prices are agreed with each partner.

Can't find your answer? Contact ai@kompa.ai.

Ready to send your first request?

Sign up, top up via VietQR and change your base_url. That's it.