For teams shipping AI in production

Ship the AI feature.
Not the infrastructure.

One Metriqual key gives your team every AI modality — text, images, audio, video, music, voice cloning — across OpenAI, Anthropic, Gemini, Mistral, and MiniMax. Automatic failover. Per-key spend controls. Full observability. One integration, production-grade from day one.

Get started free Book a demo

OpenAI · Anthropic · Gemini · Mistral · MiniMaxDrop-in OpenAI-compatible APIRust-native coreSub-5ms overhead

integration.ts

// Replace five provider integrations with one

const client = new OpenAI({

apiKey: process.env.MQ_KEY,

baseURL: "https://api.metriqual.com/v1",

});

// Text — GPT-5, Claude, Gemini, Mistral, MiniMax

// with automatic failover across all five

await client.chat.completions.create({

model: "gpt-5.1", messages,

});

// Video — Sora or MiniMax, same interface

await metriqual.video.generate({ model: "sora", prompt });

// Voice cloning — MiniMax

await metriqual.tts.create({ voice: "cloned-voice-id", text });

// Realtime WebSocket — OpenAI

await metriqual.realtime.connect({ model: "gpt-4o-realtime" });

Request stream — all modalities, one key

LIVE

The problem you stop paying for

What your engineering team stops building.

Every AI modality your product needs lives behind a different provider, a different SDK, a different auth system, a different billing account. Building and maintaining that infrastructure takes weeks of engineering — weeks your team should be spending on your actual product. And when any provider goes down, your product goes with it.

openai-sdkauth · billing · rate limits

anthropic-sdkauth · billing · rate limits

google-genaiauth · billing · rate limits

mistralaiauth · billing · rate limits

minimax-apiauth · billing · rate limits

collapses to

One key. Everything.

MQ_KEY=mq_live_••••••••••••

Every modality, every provider, automatic failover, unified billing and observability — behind one OpenAI-compatible endpoint.

→ velocity

Faster to market

Every provider integration your team doesn't build is a week spent on your actual product. One Metriqual integration instead of five. Ship the AI feature in days, not weeks.

→ reliability

Built-in reliability

When OpenAI goes down, Claude takes over automatically. Configurable failover chains per key. Your product keeps running. Your users never know.

→ cost control

Predictable costs

Per-key spend caps. Real-time cost attribution per request, per provider, per modality. Know what your AI costs before the bill arrives.

Capability

Every AI modality. One integration.

The full OpenAI generation stack, the full MiniMax generation stack, and chat routing across Anthropic, Gemini, and Mistral — unified under one key.

Capability	OpenAI	Anthropic	Gemini	Mistral	MiniMax
Chat + Streaming	✓	✓	✓	✓	✓
Function Calling	✓	✓	✓	✓	✓
Automatic Failover	✓	✓	✓	✓	✓
Embeddings	✓	✕	✕	✕	✕
Image Generation	✓DALL·E	✕	✕	✕	✓
Video Generation	✓Sora	✕	✕	✕	✓
Image-to-Video	✕	✕	✕	✕	✓
Text-to-Speech	✓	✕	✕	✕	✓
Voice Cloning	✕	✕	✕	✕	✓
Voice Design	✕	✕	✕	✕	✓
Music Generation	✕	✕	✕	✕	✓
Lyrics Generation	✕	✕	✕	✕	✓
Transcription	✓Whisper	✕	✕	✕	✕
Translation	✓	✕	✕	✕	✕
Realtime WebSocket	✓	✕	✕	✕	✕

The claim, precisely: no other managed gateway unifies the full OpenAI generation stack with the full MiniMax generation stack under one OpenAI-compatible key with per-key controls, automatic failover, and full observability.

Production controls

Not a prototype tool.

Everything you'd otherwise build yourself before letting a key anywhere near production traffic.

Per-Key Controls

Spend caps, model allowlists, key expiry, key modes. Give a key to a teammate, embed it in a product, or hand one to a client with a hard budget limit. Full control over what each key can do and spend.

⛨

PII Detection & Redaction

Regex-based content filtering per key. Emails, phone numbers, custom patterns. Toggle on/off without deleting. Live test endpoint.

Prompt Hub

Create, version, and attach system prompts to keys. Auto-injected on every request. Change the prompt in the dashboard. No code change. No deploy.

⇌

A/B Routing Experiments

Split traffic across models or providers. Collect per-request feedback. Measure what actually performs before committing.

◉

Full Observability

Cost, latency, tokens, errors — per key, per provider, per modality. Replay any request. Export logs. Webhooks for billing and usage events.

↯

Automatic Failover

Circuit breaker per provider, configurable chains per key. Primary fails, fallback fires, request completes. Your monitoring finds out — not your users.

Developer experience

One integration. Any language.

Standard OpenAI client. No proprietary SDK. No learning curve. Point your existing client at Metriqual and keep shipping.

◆Node.js SDK — typed, full coverage

◆Python SDK — full coverage

◆CLI — auth and key management from terminal

◆OpenAPI spec at /openapi.json

◆Multimodal playground — test every modality in-browser

◆Provider setup wizard with live connection test

◆Generated code snippets in Node, Python, cURL

Pricing

Start free. Scale when it works.

Every paid plan includes the full multi-provider stack. You're paying for volume and controls — not for unlocking modalities.

Free

Limited

→Playground access
→1 provider

Get started

Starter

$99/mo

50k requests / mo

→3 keys
→Spend caps
→Analytics

Start with Starter

Most teams start here

Growth

$499/mo

500k requests / mo

→Unlimited keys
→Fallbacks
→Guardrails
→Team

Start with Growth

Scale

$1,500/mo

High volume

→SLAs
→Priority support

Talk to us

Enterprise

Custom

Unlimited

→Annual contracts
→Dedicated support

Contact sales

Design partner offer — first 3 months at $99/mo, locked, regardless of plan. In exchange: honest feedback and a testimonial if it works for you. The founder personally helps you integrate.

Claim a design partner slot

One integration.
Everything your AI product needs.

Text, images, audio, video, music, voice cloning. Automatic failover. Per-key controls. Full observability. Production-grade from day one.