Developer platform

Our API · AetherGate

Use AetherGate inside your software. Create an AetherGate key here; connect OpenAI and Anthropic separately under Provider Accounts. You never need to paste provider secrets into your app's public frontend.

1 · Connect providers

Manage your OpenAI and Anthropic accounts in Provider Accounts.

2 · Create our API key

Issue a separate, revocable AetherGate key below.

3 · Start sending requests

Call our gateway or submit an asynchronous Foreman mission.

Create AetherGate key

Workspace: Loading · Create a separate key for each app or test suite.

Your keys

No API keys found for this workspace.

Building a voice receptionist?

Browse conversational voice, streaming speech and live transcription models. Audio sessions use native provider APIs; AetherGate currently routes the text reasoning stage. Native audio is billed and controlled at the provider.

Explore live voice models →

Discovery: GET /api/v1/voice/models or GET /api/v1/models?category=voice. Responses declare integration and proxy support for every entry.

Quickstart

Store your AetherGate key in a server-side environment variable called AETHERGATE_API_KEY. Never commit it or expose it in browser JavaScript.

1. Gateway · synchronous model response

curl https://aethergate.site/api/v1/chat/completions \
  -H "Authorization: Bearer $AETHERGATE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"openai/gpt-4o","messages":[{"role":"user","content":"Hello"}]}'

2. Foreman · asynchronous multi-model mission

POST /api/v1/missions with task, 2–4 connected exact model IDs, explicit model_pins, reasoning_effort=low, max_output_tokens, temperature, system_instruction, memory, and spend_cap_usd. A successful launch returns HTTP 202 and a mission ID. Poll GET /api/v1/missions?id=YOUR_ID with the same key. Exact provider model IDs and upstream pins must match your workspace configuration.

Billing: AetherGate currently charges for subscription access through Stripe. Provider token charges are billed by your connected AI providers. The displayed cost is tracked provider usage, not an additional pay-as-you-go AetherGate invoice. Separate metered API billing is not active yet.