OpenAI-compatible AI gateway

One API for every AI model you use

Use GPT, Claude, Gemini and more through one OpenAI-compatible API, or talk to all of them in live chat. One wallet, keys you control, a report for every request and automatic failover.

Works with OpenAI SDKs Live chat included Free models included
APIEX.CLOUD Example
POST /v1/chat/completions 200 OK
curl https://apiex.cloud/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"model": "openai/gpt-5.6-terra-pro",
       "messages": [...]}'
Your app One base URL
Gateway Keys, limits, billing
Upstream With failover
474Models
30Providers
5Model types
Pay as you go
Create API keys
Live chat
Models from the providers you already know
OpenAI
Qwen
Google
Mistral AI
Anthropic
DeepSeek
Z.ai
Meta
xAI
MiniMax
Moonshot AI
NVIDIA
Microsoft
ElevenLabs
Amazon
Tencent
ByteDance
Alibaba
Cohere
Voyage AI
Black Forest Labs
Xiaomi
Kuaishou
Inception
OpenAI
Qwen
Google
Mistral AI
Anthropic
DeepSeek
Z.ai
Meta
xAI
MiniMax
Moonshot AI
NVIDIA
Microsoft
ElevenLabs
Amazon
Tencent
ByteDance
Alibaba
Cohere
Voyage AI
Black Forest Labs
Xiaomi
Kuaishou
Inception
Why this gateway

Everything between your code and the model

Authentication, billing, limits, routing and reporting are handled for you, so adding AI to a product needs just one integration instead of several.

Live chat with every model New

Instead of paying a different platform for each model, use live chat with all paid and free models in one place and switch between them whenever you like. Every conversation and its history are saved, and everything is billed to the same wallet.

OpenAI-compatible API

Use GPT, Claude, Gemini and more through one base URL. Point any OpenAI SDK or HTTP client at it: chat completions with tools, JSON mode and vision input, plus embeddings and a live model catalog.

Pay only for what you use

A prepaid wallet works across every model. Before each request the gateway holds its worst-case cost, then charges the real cost and releases the rest.

Automatic failover

A model can be backed by several upstream services. If one fails, the same request moves to the next, and a circuit breaker steers traffic away from services that keep failing.

Keys you can really control

Give every project its own key with a spend limit, an expiry date, an IP allow-list, a model allow-list and a per-minute rate limit. Create several at once, revoke instantly.

Usage you can audit

Every request is logged with tokens, cost, latency and status. Charts, filters and a full report for each call show exactly where your balance went.

Safeguards against surprises

Output length is always bounded, daily spend caps and concurrency limits protect your balance, and an Idempotency-Key makes retries safe and free.

Free tier and rewards

Try free models without topping up, collect daily check-in rewards, redeem codes, and earn wallet credit when people you invite top up.

Built for a global audience

The whole site and panel work in Persian, English and Russian, with right-to-left support, light and dark themes, and optional Telegram notifications.

474AI models
30Model providers
6API endpoints
3Interface languages
Free tier

Start free — no wallet balance needed

Your allowance renews automatically every day. Need more headroom? Top up your wallet or pick a plan at any time.

Browse free models
6 Free models
10 M Free tokens per day
500 Free requests per day
6 requests / min Maximum request speed · up to 100 requests in total per hour
Live chat

Chat live with every model, paid and free, in one place

You do not need a separate subscription for every model. Open the live chat in your panel, pick any model and switch between them as you go, all paid from the same wallet.

  • One model picker for paid and free models, so you never hunt for another platform
  • Every conversation and its history are saved: search, pin, rename and continue at any time
  • File attachments, image generation, speech, transcription and video where the model supports them
  • Custom instructions and reasoning control per conversation, when the site enables them
Start live chat
For developers

Change the base URL. Keep your code

If your application already talks to an OpenAI-style API, switching takes two lines: the base URL and the key. Then call GPT, Claude, Gemini and more the same way. Everything else, from streaming to tool calls, keeps working as you expect.

  • Chat completions with tools, JSON mode, vision input and streaming
  • Embeddings and a model catalog that lists only what your key may use
  • A cost-estimate endpoint and a usage endpoint for your balance
  • Rate-limit headers, Retry-After and idempotent retries
  • OpenAI-style error responses your SDK can parse
Explore the API reference
curl https://apiex.cloud/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-5.6-terra-pro",
    "messages": [{ "role": "user", "content": "Hello!" }]
  }'
from openai import OpenAI

client = OpenAI(
    base_url="https://apiex.cloud/api/v1/",
    api_key="YOUR_API_KEY",
)

reply = client.chat.completions.create(
    model="openai/gpt-5.6-terra-pro",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(reply.choices[0].message.content)
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://apiex.cloud/api/v1/",
  apiKey: "YOUR_API_KEY",
});

const reply = await client.chat.completions.create({
  model: "openai/gpt-5.6-terra-pro",
  messages: [{ role: "user", content: "Hello!" }],
});
console.log(reply.choices[0].message.content);
$ch = curl_init( "https://apiex.cloud/api/v1/chat/completions" );
curl_setopt_array( $ch, array(
    CURLOPT_RETURNTRANSFER => true,
    CURLOPT_POST           => true,
    CURLOPT_HTTPHEADER     => array(
        'Authorization: Bearer YOUR_API_KEY',
        'Content-Type: application/json',
    ),
    CURLOPT_POSTFIELDS     => json_encode( array(
        'model'    => 'openai/gpt-5.6-terra-pro',
        'messages' => array( array( 'role' => 'user', 'content' => 'Hello!' ) ),
    ) ),
) );
$reply = json_decode( curl_exec( $ch ), true );
echo $reply['choices'][0]['message']['content'];
Reliability and control

Built so one bad request never becomes a bad bill

Requests are checked, priced and protected before they leave the gateway, and routed around failing providers while they are in flight.

Moving to the next upstream if one fails…

  • A model can be served by several services in priority order. If one fails, the identical request moves to the next, and your client never notices.
  • A service that fails 5 times within 60 seconds is skipped for 30 seconds, then probed with a single request.
  • Only failures that say something about a service's health, such as timeouts, server errors and rate limits, count towards the breaker. Mistakes in your own request never do.
  • A total time limit covers all attempts of one request, so a slow service cannot hold it forever.
Start now
Get started in minutes

From sign-up to first response

1

Create an account

Sign up with your email or continue with Google. It takes a few seconds.

2

Start free or add credit

Use free models right away, or top up the wallet or subscribe to a plan for everything else.

3

Create an API key

Name it, set the limits you want and copy the secret. It is shown only once.

4

Call the API or chat live

Send a request from any OpenAI-compatible client, or simply open the live chat and talk to any model.

Your control panel

A dashboard that answers the questions you actually have

Where did my balance go? Which key made that request? Which model is cheapest for this job? It is all one click away.

Overview

Balance, spending, requests and tokens over the period you choose, with charts.

Live chat

Chat live with every paid and free model in one place, with saved conversations, attachments, images, speech, transcription and video.

API keys

Create, limit, edit and revoke keys, one at a time or in batches.

Models

Search every model with prices, context length and capabilities, and open a page for each.

Wallet

Top up, redeem codes, see every transaction and share your referral link.

Plans

A paid plan gives you credit equal to its price. It is spent only from the plan, separate from your wallet, and expires with the plan.

Activity

Insights on how each model and key is used, so you can spot trends early.

Usage logs

Every request with its cost, tokens and duration, and a full report one click away.

Support

Open a ticket, follow the conversation and get notified when we reply.

Wallet and rewards

Add credit your way. Earn some on the way

Your balance can be shown in Toman or Dollars. Dollar amounts are converted at the exchange rate set by the site, and the real cost of each request is always recorded.

Instant card payment

Pay through the Zarinpal gateway and your wallet is credited as soon as the payment is confirmed.

Card-to-card transfer

Transfer to our card, attach the receipt, and your wallet is credited once it is approved.

Cryptocurrency

Pay from any wallet or exchange with the coins we accept. No account is needed on our side.

Plans

A paid plan gives you credit equal to its price. It is spent only from the plan, separate from your wallet, and expires with the plan.

Redeem codes

Enter a code and the credit is added to your wallet immediately.

Daily check-in

Check in every day to earn wallet credit, with bigger multipliers as your streak grows.

Referral rewards

Earn 10% of every top-up made by people you invite, credited to your wallet.

Telegram notifications

Link your Telegram account to get alerts about credit, low balance and announcements.

Simple pricing

Plans that grow with you

Choose a plan for a fixed period of credit and model access, or use the wallet for flexible pay-as-you-go usage. Plan credit and wallet money are separate.

Gemini

1 month
$5,93
+5% more usage
  • 30 days duration
  • Pay $5,93, use $6,24 worth.
Plan details
  • Credit pays only these models. Other models are paid from your wallet.
  • 5% usage discount: every 100 of real usage takes 95 from your credit.
  • Unused credit expires with the plan and cannot be refunded or moved to the wallet.
Choose plan

Gemini

3 months
$17,41
$5,80 per month +6% more usage
  • 90 days duration
  • Pay $17,41, use $18,62 worth.
Plan details
  • Credit pays only these models. Other models are paid from your wallet.
  • 6,50% usage discount: every 100 of real usage takes 93,50 from your credit.
  • Unused credit expires with the plan and cannot be refunded or moved to the wallet.
Choose plan

Gemini

6 months
$33,89
$5,65 per month +9% more usage
  • 180 days duration
  • Pay $33,89, use $37,24 worth.
Plan details
  • Credit pays only these models. Other models are paid from your wallet.
  • 9% usage discount: every 100 of real usage takes 91 from your credit.
  • Unused credit expires with the plan and cannot be refunded or moved to the wallet.
Choose plan

Chat GPT

1 month
$16,30
+5% more usage
  • 30 days duration
  • Pay $16,30, use $17,15 worth.
Plan details
  • Credit pays only these models. Other models are paid from your wallet.
  • 5% usage discount: every 100 of real usage takes 95 from your credit.
  • Unused credit expires with the plan and cannot be refunded or moved to the wallet.
Choose plan

Chat GPT

3 months
$48,15
$16,05 per month +6% more usage
  • 90 days duration
  • Pay $48,15, use $51,50 worth.
Plan details
  • Credit pays only these models. Other models are paid from your wallet.
  • 6,50% usage discount: every 100 of real usage takes 93,50 from your credit.
  • Unused credit expires with the plan and cannot be refunded or moved to the wallet.
Choose plan

Chat GPT

6 months
$93,33
$15,56 per month +9% more usage
  • 180 days duration
  • Pay $93,33, use $102,56 worth.
Plan details
  • Credit pays only these models. Other models are paid from your wallet.
  • 9% usage discount: every 100 of real usage takes 91 from your credit.
  • Unused credit expires with the plan and cannot be refunded or moved to the wallet.
Choose plan
No plan? No problem.

Pay as you go with your wallet

Skip fixed plans entirely — top up your wallet once and pay only for what you actually use, any time.

Get started free
AI Models

The full model directory

Every model we support, in one list — click any one to see its full features, pricing, and details.

gpt-5.6-luna-pro

openai/gpt-5.6-luna-pro

GPT-5.6 Luna Pro is the same underlying model as [GPT-5.6 Luna](Apiex), served with reasoning.mode set to pro for higher-qualit…

Max Context
1,1M
Input / 1M
$0,24
Output / 1M
$1,44

gpt-5.6-luna

openai/gpt-5.6-luna

GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series. It is suited for high-volume, latency-sensitive tasks…

Max Context
1,1M
Input / 1M
$0,24
Output / 1M
$1,44
No cost

Free models, ready right now

No subscription required — start building with these today.

Questions, answered

Frequently asked questions

Still unsure about something? The documentation goes deeper, and our team is happy to help.

Can I chat with the models without writing code?

Yes. The live chat in your panel lets you talk to all paid and free models in one place and switch between them whenever you like, instead of paying a separate platform for each model. Your conversations and their history are saved, and it is billed to the same wallet.

Is it compatible with the official OpenAI SDKs?

Yes. Set the SDK's base URL to the gateway address shown in the documentation and use your gateway API key. Requests and responses follow the OpenAI Chat Completions and Embeddings formats, errors use the envelope the SDKs already understand, and you can call GPT, Claude, Gemini and other models the same way.

How does billing work?

You top up a prepaid wallet and every request is charged from it. Before a request is forwarded, the gateway holds the most it could possibly cost. When the answer arrives, you are charged the real cost and the rest of the hold is released immediately. If your available balance cannot cover the hold, the request is refused with a 402 error and nothing is charged.

Can I try it without paying?

When the free tier is enabled, models marked as free can be used without any wallet balance, within a token and request allowance that renews automatically. You can also earn credit through the daily check-in, redeem codes and referrals.

Can I limit what an API key is allowed to do?

Yes. Each key can have its own spend limit (total, daily or monthly), an expiry date, an IP allow-list, a list of allowed models and a requests-per-minute limit. A key can be revoked at any time and the change applies immediately.

What happens if an upstream provider is down?

A model can be served by more than one upstream service. When the first one fails, the gateway retries the identical request on the next one without your client noticing. A circuit breaker temporarily skips a service that keeps failing, so one outage does not slow down every request.

Does streaming work?

Yes. Send "stream": true and read standard server-sent events ending with [DONE]. Where live streaming is enabled, tokens are forwarded as they are generated and the wallet is settled when the stream ends; otherwise the finished answer is delivered in the same event format.

How can I protect myself from accidental double charges?

Send an Idempotency-Key header with your request. If the same key is sent again with the same body within 24 hours, you receive the original response and are not charged a second time.

Can I see what each request cost?

Yes. Your panel lists every request with its model, tokens, cost, duration and status, and each row opens into a full report. You can also ask the API itself for your balance and usage, or estimate the maximum cost of a request before sending it.

Do my wallet credits expire?

Money added to the wallet by a top-up does not expire. Plan credit is different: it lasts as long as the plan and then burns. It cannot be refunded or moved to the wallet.

Which languages are supported?

The site, panel and documentation are available in Persian, English and Russian, and you can switch language at any time from the header.

Ready to build with one API?

Create an account, make a key and send your first request in a few minutes.