O
AETHERAROUTER v2.1
LIVE CATALOGOpen workspace ↗
AETHERA ROUTER • GLOBAL CONTROL

One route
for every model.

Aethera gives your team one OpenAI-compatible gateway for model routing, transparent usage, and budget-aware execution.

01 Stable aliases02 Provider-aware routes03 Immutable ledger
ROUTER / LIVE VIEW OPERATIONAL
ACTIVE PUBLIC ALIAS

Loading catalog…

01 / —
REQUESTchat.completionsOpenAI-compatible
→
DECISION LAYERAETHERAhealth · budget · route
→
PROVIDER ROUTELive model routepricing from catalog
STATUSREADY
PRICINGUSD / 1M
ACCESSWORKSPACE
LIVE—ACTIVE MODELS
POSTapi.aethera.com/v1/chat/completions
BUILT FOR TEAMS THAT NEED TO KNOWwhat ranwhere it ranwhat it costwhat happens next

THE CONTROL SURFACE

Model access is easy.
Model confidence takes a layer.

Keep your application integration simple while Aethera handles the operational details.

01

Route with intent

Expose stable public aliases while Aethera selects enabled provider routes and keeps provider identity visible.

Explore models ↗
02

See the real cost

Reserve against request budgets, settle from reported usage, and keep wallet movements in the ledger.

Open usage ledger ↗
03

Ship without guesswork

Test requests in the playground, inspect responses, then create an API key for your integration.

Start in playground ↗

MODEL CATALOG

Choose a route.
Keep the details visible.

Select an active Aethera alias. Inspect provider, limits, pricing, and availability.

Open full catalog ↗

ACTIVE MODELS

Loading…
01–04

Select a model
to inspect its route.

PUBLIC MODEL ALIAS

Loading catalog…

Aethera•OPERATIONAL

01 / —
AVAILABLEUSD / 1M TOKENS

Live public model alias routed through an active provider route.

CONTEXT WINDOW—
MAX OUTPUT—
INPUT / 1M—
OUTPUT / 1M—

Loading live model catalog…

TRANSPARENT PRICING

Pay for recorded usage.
No invented plan tiers.

Model rates are shown per one million tokens. Your request ledger records actual charges, while prepaid credits remain under your control.

PREPAID CREDITS

Payment processing is not connected yet. Top-up requests require manual payment verification; credits are not added until confirmed.

THE REQUEST PATH

Simple at the surface.
Disciplined underneath.

01

Authenticate

Use a workspace session or API key.

02

Choose an alias

Keep your app connected to a stable public model ID.

03

Execute safely

Routing, reservation, provider call, and settlement stay observable.

04

Read the result

Response and financial record arrive with the request trail.

FOR DEVELOPERS

Your first request
should feel familiar.

Open the workspace, test a model in the playground, and move to an API key when your integration is ready.

Open developer portal →
request.sh
curl /v1/chat/completions  -H "Authorization: Bearer ***"  -d '{"model":"public-api","messages":[…]}'
✓ OpenAI-compatible request surface

DOCUMENTATION / V2.1

Integrate with Aethera.

Quickstart, endpoint reference, billing behavior, and error handling—all without leaving the landing page.

Open workspace ↗
01 / START HERE

Make your first request.

Open the workspace to test a model and create an API key. Keep the raw key private; it is shown only once.

request.sh
curl https://api.aethera.com/v1/chat/completions \\
  -H "Authorization: Bearer ${AETHERA_API_KEY}" \\
  -H "Idempotency-Key: replace-with-unique-request-id" \\
  -H "Content-Type: application/json" \\
  -d '{"model":"public-api","messages":[{"role":"user","content":"Hello"}]}'

Use a fresh idempotency key for each logical request. Repeating the same key and payload returns its stored result; reusing the key with a different payload returns HTTP 409. Keys are scoped to the authenticated user.

02 / ACCESS

Authenticate once.

Send your API key as a Bearer token to public API endpoints. The authenticated workspace uses its browser session. Never place a key in client-side browser code.

HEADERAuthorization: Bearer ${AETHERA_API_KEY}

Never expose an API key in browser code or commit it to source control.

03 / CATALOG

Choose a public alias.

List enabled public aliases before making a request. Pricing and availability are controlled by the active server catalog.

04 / REQUESTS

Chat completions.

Send a JSON body with a public model alias and messages. Include a unique Idempotency-Key header for each logical request; the endpoint rejects a missing key with HTTP 400.

POST/v1/chat/completions
payload.json
{
  "model": "public-api",
  "messages": [{"role": "user", "content": "Your prompt"}],
  "temperature": 0.7
}
05 / ACCOUNTING

Usage becomes a record.

Each completed request records usage and its wallet settlement in the request history and ledger. USD is the canonical billing currency.

06 / RECOVERY

Handle errors clearly.

Errors use JSON responses with a useful detail or normalized error message. Check the HTTP status and do not charge a failed request client-side. Request history and the ledger in the workspace are the source of truth.

When a request fails, inspect the response status and error payload before retrying. Use a new idempotency key only for a new logical request.