Your cookie choices

We use necessary cookies to make this site work. With your permission, analytics cookies help us understand usage and marketing cookies support advertising.

You can reject non-essential cookies and still use the site. Change your choices anytime in Cookie settings.

Cookie settingsCookie policy

LLM Gateway review

Updated September 28, 2026 · DeployCue Editorial Team · Facts checked September 26, 2026

We earn a commission when you sign up through links on this site. Partner status can affect where a brand appears in our rankings. Partners with a live offer in your country come first, then every other brand by score.

Our verdict

LLM Gateway suits teams that already hold provider keys and want a single OpenAI-compatible endpoint across 46 providers, since bring your own key and self-hosting carry no platform charge. Teams who prefer a predictable subscription bill should weigh Salad first, since the approaches settle cost differently.

Key facts

Best for
One OpenAI-compatible endpoint over many providers, with bring your own key at no platform charge.
Fee model
Pay-as-you-go credits at provider list rates plus flat 5% platform fee; BYOK and self-hosting free; enterprise custom↗
Providers routed
46↗
Bring your own key
Yes↗
OpenAI-compatible API
Yes↗
Free tier
Yes↗
Legal entity
Polar Lights LLC↗
Headquarters
United States↗
Support channels
Email, Chat, Help-centre, Community↗
Official website
llmgateway.io

Facts checked September 26, 2026, 4 sources

Apps and profiles

Elsewhere
Xgithub.com

Pros and cons

Pros

  • Routes across 46 providers behind a single OpenAI-compatible API, so model swaps stay inside the same endpoint.
  • Bring your own key is supported and listed as free, so existing provider contracts and negotiated rates carry over.
  • Self-hosting is free as well, which suits teams that want the routing layer inside their own network.
  • A free tier lets a small team test routing and key handling before any spend.
  • Support runs through email, chat and help centre, pairing self-serve reading with email and chat.

Cons

  • The platform fee applies on top of provider bills, so LiteLLM's self-host free forever route is the cheaper gateway layer for teams able to run it.
  • Support runs on email, chat and help centre only, with no phone line listed for incidents.
  • Enterprise terms in the fee model are custom, so large deployments are quoted rather than listed at a published price.
LLM Gateway

Ready to try LLM Gateway?

Scored 9.2 / 10 by DeployCue.

See alternatives

Who it suits

  • Teams that already hold provider keys and want a single OpenAI-compatible endpoint across 46 providers.
  • Small teams that want a free tier to test routing before committing spend.
  • Organisations that prefer to self-host the gateway and keep billing with their existing providers.

Who should look elsewhere

  • Teams wanting a flat monthly subscription instead of usage charges, where Salad on Flat monthly subscription or pay per token reads more predictably.
  • Engineers who want a gateway layer free forever and can run open source, where LiteLLM fits better.
  • Buyers who need a published enterprise price before a first call, since enterprise terms are custom.

What it does well

LLM Gateway sits between an application and the model providers behind it, and the fact that separates it from the gateways next to it is routing breadth held together with key control. It routes across 46 providers, and bring your own key is supported, so a team can point credentials it already holds at a single endpoint rather than opening a fresh contract with every vendor. The API is OpenAI-compatible, which means most client libraries and agent frameworks work against it with a base URL change rather than a rewrite. Together these describe a gateway built for portability: providers can change, credentials stay with the team, and the calling code keeps its shape.

Pricing

A couple of readings follow for a reader weighing cost. Teams routing on their own keys pay providers directly and avoid the platform charge, so the gateway layer costs nothing. Teams buying credits through the gateway pay provider list rates plus the flat platform fee, which keeps provider pricing visible rather than folding it into a subscription. A free tier is also offered, so a small team can start at no spend and judge routing behaviour before volume matters. Against Salad on Flat monthly subscription or pay per token the contrast is a subscription set against usage charges, and against LiteLLM's self-host free forever route it is a platform fee set against none.

What it connects to

The integration story rests on a pair of facts. The API is OpenAI-compatible, so existing client code, SDK calls and framework integrations can point at the gateway through configuration rather than a rewrite, which keeps migration cost low. Bring your own key is supported, so a team plugs in the credentials it already holds with each provider instead of storing new secrets on a gateway account, and the contracts behind those keys stay where they are. For a reader weighing gateways, the practical question is whether both the existing code and the existing keys can stay as they are, and on these facts they can.

Who runs it

The service is operated from the United States, with the headquarters country given as US, and the legal entity behind the terms is Polar Lights LLC. Support runs through email, chat and help centre, which pairs self-serve material with direct channels: a help centre and community for reading, email and chat for questions that need a reply. For a buyer these details carry practical weight, since a named legal entity gives procurement something to check and the headquarters country shows which jurisdiction the contract sits under. A phone channel is not among email, chat and help centre, which is worth weighing for anyone running a gateway as an on-call dependency.

How LLM Gateway compares

Frequently asked questions

Is there a free tier?+

Yes, a free tier is offered, which matches most gateways in this comparison (81.7%). Teams can test routing and key handling at no spend, and teams already holding provider keys can skip the credits path altogether since bring your own key is free.

Which providers can it route to?+

It routes across 46 providers, and that breadth matters for failover, since a provider can change without the calling code moving. On providers routed it ranks 3 among 17 brands, so the list is broad without being the broadest in the set.

Can I use my own provider keys?+

Yes, bring your own key is supported, and most gateways here do the same (70%). The fee model lists bring your own key as free, so a team can keep its existing provider contracts and negotiated rates while pointing the gateway at credentials it already holds.

Is the API OpenAI-compatible?+

Yes, the API is OpenAI-compatible, which is the norm across this comparison (100%). Existing client libraries, SDK calls and agent frameworks can point at the gateway through a base URL change, so migration comes down to configuration rather than rewriting the calling code.

Who runs LLM Gateway?+

The service is operated from the United States, with the headquarters country given as US, and the legal entity named in the terms is Polar Lights LLC. That gives procurement a named counterparty and shows a buyer which jurisdiction the contract sits under.

How can I get support?+

Support runs through email, chat and help centre. A help centre and community cover self-serve questions, while email and chat cover anything that needs a reply, so reading material is there before a conversation starts. A phone line is not among the listed channels.

How does it compare with Salad on cost?+

LLM Gateway charges provider list rates plus a flat platform fee on credits, while Salad bills Flat monthly subscription or pay per token. A team with steady traffic may prefer a fixed subscription and simple budgeting, while a team with lumpy usage or its own provider keys may prefer usage charges with no platform fee.

Sources

  • llmgateway.io/contact checked September 26, 2026

    Backs: Headquarters, Support channels

    “LLM Gateway is operated by Polar Lights LLC 16192 Coastal Highway Lewes, DE 19958 United States”

  • llmgateway.io/legal/terms checked September 26, 2026

    Backs: Legal entity

    “operated by Polar Lights LLC, 16192 Coastal Highway, Lewes, DE 19958, United States”

  • llmgateway.io/pricing checked September 26, 2026

    Backs: Bring your own key, Fee model, Free tier, OpenAI-compatible API

    “Bring Your Own Keys (BYOK) - Price: free. Your provider bills you directly at its own rates.”

  • llmgateway.io/providers checked September 26, 2026

    Backs: Providers routed

    “Access 536 models from 46 leading AI providers through our unified API”

Compiled from 4 source pages, checked September 26, 2026. We did not open an account.

LLM Gateway

LLM Gateway

One OpenAI-compatible endpoint over many providers, with bring your own key at no platform charge.

9.2/10See alternatives