AI GATEWAY FOR OPENAI + CLAUDE

When your AI provider goes down, you don't.

Metriqual sits between your app and every model provider. Automatic failover, live web search with citations, private knowledge bases, voice and vision — behind one key, on the API you already use.

Start Building See it fail over

2 lines to integrate·<12ms overhead·your keys, your billing

metriqual — live trafficLIVE

YOUR ENTIRE MIGRATION

- base_url = "https://api.openai.com/v1"
+ base_url = "https://api.metriqual.com/v1"

WHAT HAPPENS WHEN A PROVIDER FAILS

$ POST /v1/chat/completions   model=gpt-4o

→ openai        connect 2.00s … timeout

↻ failover      next step in chain

→ anthropic     claude-fable-5 · 200 OK · 412ms

✓ delivered     0 requests dropped

ROUTES ACROSS OpenAI

OpenAI

Anthropic

GeminiSOON Mistral

MistralSOON MiniMax

MiniMaxSOON

One endpoint

Every provider. Every modality. One key.

Route across OpenAI and Anthropic with automatic failover — Gemini, Mistral and MiniMax land next. Text, vision, voice, live web search and private knowledge bases, all through the same endpoint.

2 linesto integrate

<12msrouting overhead

100%conversation history carried

Everyrequest logged & attributable

2 liveproviders — 3 more landing

OpenAI

Claude

Gemini

Mistral

MiniMax

OpenAI

Claude

Gemini

Mistral

MiniMax

OpenAI

Claude

Gemini

Mistral

MiniMax

OpenAI

Claude

Gemini

Mistral

MiniMax

Text & Chat

Voice + Cloning

Video / Sora

Image Generation

Music Generation

Image-to-Video

Text & Chat

Voice + Cloning

Video / Sora

Image Generation

Music Generation

Image-to-Video

Reliability

Zero dropped requests, even mid-outage.

Automatic failover reroutes every request to a healthy provider in milliseconds, through outages, rate limits, and regional degradation. A single-provider setup goes down with it.

Metriqual Failover active

Provider outage → rerouted

0failed requests

rerouted to a healthy provider mid-outage

Failover overhead

1–5ms

Providers

History carried

Yes

Single-provider setup Provider down

Provider outage → no fallback

5xxto your users

every in-flight request fails until the provider recovers

Failover

None

Providers

History carried

MCP PROCESSING ENGINE

Change two lines.
Keep everything else.

Replace your OpenAI API key and base URL. Everything else in your codebase, from prompts to parameters, remains completely unchanged.

Start integration →

agent.ts

1const client = new OpenAI({

2 apiKey: process.env.OPENAI_KEY,

3 baseURL: "https://api.openai.com/v1",

4});

✓ Build SuccessUTF-8

Auto routingLoad balancingFailover

Metriqual×

app.metriqual.com/dashboard/create

Research AgentDraft

BuilderPreviewLogsMetrics

All changes saved

Components

Trigger

Agent

Model

Tool

Data Source

Logic

Response

My Library

Saved Agents

Custom Tools

Knowledge Bases

Agent Name

Research Agent

Description

Research and analyze any topic using real-time data and internal knowledge.

Model

AI Claude Fable 5

System Instructions

Research and analyze any topic using real-time data and internal knowledge.

Configure Agent

Name

Research Agent

Description

Research and analyze any topic using real-time data and internal knowledge.

Model

Claude Fable 5

System Instructions

You are a research assistant. Use available tools to find, analyze and synthesize information accurately.

Temperature

0.2

Max Tokens

2048

Advanced Settings

Describe

Build

Deploy

Metriqual×

app.metriqual.com/dashboard/create

Research AgentDraft

BuilderPreviewLogsMetrics

All changes saved

Components

Trigger

Agent

Model

Tool

Data Source

Logic

Response

My Library

Saved Agents

Custom Tools

Knowledge Bases

100%

Trigger

Webhook / API

Research Agent

Claude Fable 5

Web Search

Brave Search

Knowledge Base

Notion

Data Source

PostgreSQL

Response

JSON / Text

Configure Agent

Name

Research Agent

Description

Research and analyze any topic using real-time data and internal knowledge.

Model

Claude Fable 5

System Instructions

You are a research assistant. Use available tools to find, analyze and synthesize information accurately.

Temperature

0.2

Max Tokens

2048

Advanced Settings

Describe

Build

Deploy

Metriqual×

app.metriqual.com/dashboard

Research AgentDraft

BuilderPreviewLogsMetrics

All changes saved

STEP 3 OF 3

•LIVE AGENT BUILDER

Your agent is live!

Research Agent is deployed and ready to use.
You can start using it in your environment or share it with your team.

Deployed successfully

✦

Research AgentDeployed

Claude Fable 5 • Deployed just now

Agent ID: agt_01H6X3Q7W2FZBMBV4B6E2P

Overview

Research and analyze any topic using real-time data and internal knowledge.

◈

Tools

▣

Data Sources

□

Knowledge Base

◷

Response Format

Connected Components

Trigger

Webhook / API

Research Agent

Claude Fable 5

Web Search

Brave Search

Knowledge Base

Notion

Data Source

PostgreSQL

Response

JSON / Text

Deployment Details

Environment

Production

Deployed On

May 24, 2025, 10:42 AM

Deployed By

mql-user@example.com

Version

v0.1.0

Test your agent

Ready

Analyze the latest AI market trends and summarize key insights.

Response

Completed in 4.2s

Generative AI adoption is accelerating across industries, with 78% of organizations increasing investment.

Enterprise focus is shifting from experimentation to production-scale deployments.

Key growth areas include AI agents, multimodal models, and domain-specific solutions.

Data quality and governance remain critical challenges for successful AI implementation.

Describe

Build

Deploy

Connect OpenAI, Claude, Gemini, Mistral, and MiniMax through a single endpoint. Manage routing, reliability, and usage without maintaining multiple AI integrations.

01 / 03

Live web search

One flag and your agent answers with current information — and cites its sources. Works on OpenAI and Claude, through the same endpoint.

Citations on every answer

01 / 05

How we compare

Everything a gateway should do, in one place.

	Metriqual	Direct provider	Portkey	LiteLLM
Automatic failover across providers
2-line integration (no SDK rewrite)
Conversation history carried on failover
Prompt versioning & hub
Per-agent spend controls
Multi-modal (voice, video, image, music)
Fully managed, nothing to host
Unified observability

partial / limited

PRICING

Built for every stage.

Free

EVALUATE & PROTOTYPE

1 agent endpoint
10K requests / month
3 providers
No automatic failover
Community support

GET STARTED

Builder

SHIP YOUR FIRST AI FEATURE

$0/month

5 agent endpoints
100K requests / month
All 5 providers
Automatic failover
Prompt Hub + versioning
30-day logs
Email support

START BUILDING

Recommended

Growth

PRODUCTION AT REAL USAGE

$0/month

Unlimited agent endpoints
500K requests / month
Per-agent spend controls
PII redaction
Webhooks + A/B routing
Team access (5 seats)
Priority support

START FREE TRIAL

Scale

AI AS CORE INFRASTRUCTURE

$0/month

Unlimited everything
2M+ requests / month
Conversation history SLA
Custom fallback logic
Datadog/Splunk export
1-year retention
Dedicated Slack

GET STARTED

Enterprise

DOWNTIME = REVENUE IMPACT

Custom

Everything in Scale
Custom SLAs with penalties
Dedicated solution engineer
Security & compliance review
Annual contract
Co-authored case study

CONTACT SALES

Need a custom deployment?

For enterprise teams needing custom SLA targets and private gateways.

Enterprise SLAsDedicated infrastructurePriority engineering support

Talk to us→

FAQ

Frequently asked questions.

Everything developers need to know before building production-ready AI agents.

Your AI agent is live.

Build, evaluate and deploy production-ready AI agents from one platform.
Your next deployment is just one click away.

Book a Call

✓ Production Ready99.99% UptimeSub-12ms Overhead

When your AI provider goes down, you don't.

Every provider. Every modality. One key.

Zero dropped requests, even mid-outage.

Change two lines. Keep everything else.

Your agent is live!

Live web search

Everything a gateway should do, in one place.

Built for every stage.

Free

Builder

Growth

Scale

Enterprise

Need a custom deployment?

Frequently asked questions.

Your AI agent is live.

Change two lines.
Keep everything else.