NOW IN PUBLIC BETA
The inference layer for teams that ship.
Deploy, route, and observe any model — open or frontier — behind a single API. Sub-50 ms routing, usage-based pricing, and zero idle cost.
99.99% uptime / 41ms p50 routing / 120+ models / SOC 2 Type II
POWERING INFERENCE FOR FAST-MOVING TEAMS
NORTHWIND
Lumen
Vertex
OBSIDIAN
Halcyon
Quanta
Meridian
Stric
WHY AETHER
The boring parts of inference, handled.
Adaptive routing
Every request lands on the cheapest model that clears your quality bar.
41ms p50
Edge-cached weights keep cold starts off your critical path.
120+ models
Open weights and frontier APIs behind one schema.
Trace every token
Per-request cost, latency, and prompt diffs on one dashboard.
Compliant by default
SOC 2 Type II, zero-retention routing, regional pinning.
Scales to zero
No idle GPUs. You pay for tokens, never warm capacity.
Point your SDK at Aether. Keep the rest.
Drop-in compatible with the OpenAI client. Change the base URL, keep your code.
8.4B
tokens/day
41ms
p50
62%
avg cost cut
route.ts
import OpenAI from “openai”
const client = new OpenAI({
baseURL: “https://api.aether.ai/v1”,
apiKey: process.env.AETHER_KEY,
})
await client.chat.completions.create({
model: “aether/auto”,
messages,
})
Pay for tokens. Never for idle.
Hobby
$0 forever
✓ 1M tokens/mo
✓ 20 models
✓ Community support
Pro
POPULAR
$49/mo
✓ Usage-based after 50M
✓ All 120+ models
✓ Priority routing
✓ Full observability
✓ SOC 2 report
Scale
Custom
✓ Volume pricing
✓ Dedicated capacity
✓ SLA + SSO
Stop babysitting your inference stack.
Route every model through one fast, observable layer — and spend your time shipping product instead.
Start building free