Router by Mega Promoting · AI Gateway

The AI gateway for companies across Eastern Europe.

OpenAI-compatible API to dozens of GPT, Claude, Gemini, image, audio, and realtime models. Keys, budgets, billing in MDL and EUR — all from a single console.

Create account See live models

live models

99.9%

SLA uptime

doc pages

setup

openai-router.ts

1import OpenAI from "openai";
2 
3const client = new OpenAI({
4  apiKey: process.env.ROUTER_API_KEY,
5  baseURL: "https://api.megapromoting.com/v1",
6});
7 
8const res = await client.chat.completions.create({
9  model: "gpt-5.4-mini",
10  messages: [{ role: "user", content: "Salut!" }],
11});
12 
13console.log(res.choices[0].message.content);

Platform live

today: 0·last hour: 0

AI tokens served through Router by MP — cumulative since launch

Top models today

live · 30s

2026 model news

Hand-curated by the Mega Promoting team. Launches, betas, pricing changes.

See all live models →

1 iun 2026STABLE

GPT-5.5

Azure OpenAI

GPT-5.5 officially hits mainstream on Azure with pricing down ~18% and a 1M-token context. Available on the gateway.

15 mai 2026BETA

Claude Opus 4.8

Anthropic

Opus 4.8 public preview: extended reasoning, improved vision, and parallel tools. Accessible via Router keys.

1 mai 2026STABLE

Gemini 3.1 Pro

Google

Gemini 3.1 Pro: 2M-token context, native audio + video understanding, and a Live API stable for production.

10 apr 2026NEW

Sora-2

Azure OpenAI

Video generation up to 60s, 1080p, synced audio. Available through Router on Business and Enterprise plans.

20 mar 2026STABLE

GPT-5.4-pro reasoning

Azure OpenAI

A reasoning-dedicated variant with extended chain-of-thought, 78% SWE-bench score. Recommended for coding agents.

15 feb 2026STABLE

gpt-realtime-15

Azure OpenAI

Voice-to-voice with <300ms latency, 24 native voices, and tool-call support. Used by Kallina and call centres.

Why Router by MP

Operational control, not just a proxy.

Router by MP is built for everyday work. Not “yet another OpenAI wrapper”.

Per-project API keys, fully controlled

Create separate keys with budgets, lifetimes, and allowed models. The token is shown once; afterwards only the fingerprint remains. Rotate in 30 seconds.

Local billing in MDL and EUR

Wallet, Stripe top-up, Moldova e-Factura, and Enterprise SLA — all from a single console.

Cost tracking: 1h / 24h / 7d / 30d

Spend per key and per model, spike warnings at 70/90%, and a monthly report for accounting.

OpenAI SDK compatible

Just set baseURL. Works with Node, Python, .NET, Go, LangChain, LlamaIndex.

Support in Romanian, Russian, English

Implementation, debugging, and architecture discussed directly with the team — no ticket queues.

Transparent markup

$X in your wallet = $X of AI usage at the listed prices. No hidden fees.

Real-world use cases

From chatbot to voice agent.

See all use cases

Chatbots

Support widgets, lead qualification, and CRM replies with a separate budget per project.

AI agents

Code, research, and ops agents with a whitelist of allowed models per key.

Content

Rewriting, scoring, editorial planning, and translations for local teams.

Image generation

Controlled access to gpt-image-2 whenever the model is exposed on your key.

Voice & realtime

Voice assistants and PBX bridges through the available realtime models.

Embeddings & RAG

Indexing, semantic search, and knowledge bases with separate tracking.

3 lines of code

Swap the baseURL and you’re done.

Keep using the OpenAI SDK exactly as you do today. Same code in Python, Node, .NET, Go. The difference? You control the key, the budget, and the exposed models — we handle billing and compatibility.

Streaming, tools, vision, embeddings — just like OpenAI
Automatic fallback between models on rate-limits
Full request/response logging on your key (admin only)

Read the docs SDK integrations

streaming.py

1# Python — acelasi cod ca pe OpenAI
2from openai import OpenAI
3 
4client = OpenAI(
5    api_key=os.environ["ROUTER_API_KEY"],
6    base_url="https://api.megapromoting.com/v1",
7)
8 
9stream = client.chat.completions.create(
10    model="claude-sonnet-4.6",
11    messages=[{"role": "user", "content": "..."}],
12    stream=True,
13)
14for chunk in stream:
15    print(chunk.choices[0].delta.content or "", end="")

Pricing

Simple. Transparent.

Pay-as-you-go from $5. Or a subscription with monthly credit included.

Free

Pay-as-you-go

All models
$1 credit at signup
Full documentation

Start for free

Pro

$20

≈ 352.18 MDL at the central-bank rate

/month · $20 monthly credit

Priority queue
Custom prompts
Telegram support

Choose Pro

Business

$80

≈ 1.409 MDL at the central-bank rate

/month · $80 monthly credit

Automatic e-Factura
99.9% SLA
Volume discounts

Choose Business

See all plans + per-model pricing

Frequently asked questions

In short, how it works.

Does it work with my existing OpenAI SDK?

Yes. The only change is the `baseURL`. The rest of your code stays identical in Node, Python, .NET, or any language with OpenAI support.

How is pricing calculated?

Per-model rates (input / output / image / audio). $1 in your wallet = $1 of usage at the prices listed in /pricing. No surprises.

Can I issue a fiscal invoice in Moldova?

Yes — we integrate with e-Factura (mfinante.gov.md) on the Business and Enterprise plans. 0% VAT on service exports.

Which models are available?

GPT, Claude, Gemini, image models (gpt-image-2), embeddings, speech/audio, and realtime — depending on availability. Live list in /models.

Do you offer an SLA and dedicated support?

The Business plan includes a 99.9% SLA + Telegram support. Enterprise comes with a negotiated SLA, a dedicated account manager, and automatic e-Factura.

Can I set per-key limits?

Yes. On each key: USD budget, period (daily/monthly/total), model whitelist, and spike warnings at 70% and 90%.

Shall we start today?

$1 credit at signup — no card required. Connect the OpenAI SDK in 3 minutes.

Create a free account Talk to us