Looking Beyond Microsoft Foundry?

Foundry gives you the models Azure hosts — after you create resources, deployments, and quotas. LLM Gateway gives you multiple labs and clouds — including Azure itself — behind one open-source, OpenAI-compatible API. Managed access for supported routes.

Cloud-NeutralOpen-Source CoreManaged Model AccessAzure Built In
Start for Free
View Documentation
Compare platforms

Compare cross-provider routing and Azure-native tooling

Compare LLM Gateway and Microsoft Foundry (formerly Azure AI Foundry) features side by side

Why choose LLM Gateway?

No cloud lock-in — Azure is one of the supported providers behind a single OpenAI-compatible API
Managed model access — use credits for supported routes; Azure BYOK still follows deployment quotas
Open-source core — self-host under AGPLv3; enterprise features have separate terms
Keep your Azure setup — bring your Azure credentials and route through Foundry with 0% markup

LLM Gateway

OPEN-SOURCE & CLOUD-NEUTRAL

From $0

Free core; hosting costs separate

Microsoft Foundry

FORMERLY AZURE AI FOUNDRY

Usage-based

Requires an Azure subscription

Cloud-neutral

Choose from supported providers across clouds

Azure only

Open source & self-hostable

Run the core on your infrastructure; Enterprise has separate terms

AGPLv3

OpenAI-compatible API

Common request formats; features depend on the model

Varies by model

Setup required

What you need before your first request

Sign up, copy a key
Azure subscription, resource, deployments

Providers behind one API

Clouds, labs, and fast independent hosts

See live provider catalog
See Foundry model catalog

Frontier model coverage

Check live catalogs for the models you need

Cross-provider catalog
See Foundry model catalog

Fast inference hosts

Choose independent inference hosts

Deployment requirements

Setup depends on the billing and provider route

Managed routes; BYOK keeps provider requirements
Deployments + TPM quotas

Automatic provider routing

Routes on live uptime, throughput, price, and latency

Model router with automatic fallback

Failover across providers

Retry within the configured model and provider pool

Fallback within Foundry's supported pool

Response caching

Built-in caching for repeated requests

Redis, 10s–1yr TTL
Prompt caching (select models)

Route to Azure

Keep Azure OpenAI and Foundry models in the mix

Built-in providers
—

Bring your own keys

Use your own provider credentials

0% markup
N/A (Azure billing)

Transparent platform fee

Predictable, easy-to-reason-about pricing

5% or 0% (BYOK)
Usage-based or provisioned; service charges vary

Real-time cost analytics

Per-request cost, latency, and usage in one dashboard

Azure Monitor / Cost Management

Content guardrails

Content checks; coverage and pricing vary by product

Enterprise

Already on Azure? Keep it — and stop depending on it.

Azure OpenAI and Microsoft Foundry are built-in LLM Gateway providers. Bring your Azure credentials and your traffic keeps flowing with 0% markup — with configurable provider fallback, response caching, and cost analytics. Content guardrails require Enterprise. Check our live catalog for supported routes outside Foundry.

Start Free with LLM GatewayView Pricing Details

No credit card required • Self-host option available • Enterprise support available

FAQ

LLM Gateway vs Microsoft Foundry

Common questions about using LLM Gateway alongside or instead of Microsoft Foundry, formerly Azure AI Foundry.

Can't find an answer? Contact us

Microsoft Foundry is the current name for the platform previously called Azure AI Foundry. It combines model deployment, agents, evaluation, observability and governance. We retain the original comparison URL for existing links.