EmpryoEmpryo.beta
← Providers
provider

LLM Gateway with Empryo

LLM Gateway (llmgateway.io) is the fastest way to get Empryo running: one unified API in front of 40+ providers and 200+ models, with real-time cost analytics and automatic failover. Empryo treats it as a first-class provider with browser login built in — no key copying. Log in once and route any of its models to any task-router slot; your usage shows up in the same ledger Empryo audits its benchmarks with.

Popular models

claude-sonnet-4-6claude-haiku-4-5gpt-5.5gemini-3.1-pro200+ more

Setup

The 60-second path — browser login

No dashboard, no key copying. In the terminal:

empryo   # then type /login

/login opens your browser on the LLM Gateway dashboard, which mints a key and hands it straight back to Empryo over a local loopback — it lands in your OS keychain and models are available immediately (Ctrl+L to pick one). In the desktop app the same flow lives under Settings → Keys → LLM Gateway.

Two ways to pay — both built in

  • Account credits — pay-as-you-go top-ups (/login). You pay provider rates plus a small platform fee; no subscription.
  • DevPass subscription — a flat plan billed per request instead of per token (/login llmgateway-devpass). The desktop Keys panel shows both options side by side.

Already have a key? The classic path works too:

empryo --set-key llmgateway <your-key>

Why it pairs well with Empryo

  • One login, 200+ models — mix providers per task-router slot (spark on a cheap model, ember on a strong one) without juggling accounts.
  • Failover — if a provider degrades, the gateway reroutes; your session keeps going.
  • Cost clarity twice over — the gateway tracks spend per model, and Empryo's own ledger tracks it per session. The two should agree; that's the point.

Verify

empryo --list-providers   # llmgateway shows as configured
empryo --list-models llmgateway