Documentation

Quick start

Try chat completions with no API key on the Instant tier (rate-limited by IP):

curl https://api.eaon.dev/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4-flash",
    "messages": [{ "role": "user", "content": "Say hello in one sentence." }]
  }'

For Plus and Pro, create an API key in the dashboard and send Authorization: Bearer <eaon-api-key>.

Plans

GEThttps://api.eaon.dev/v1/plans

Chat responses include an eaon object with routing metadata and your active plan. Pro includes unlimited flagship access to gpt-5.5.

Chat completions

POSThttps://api.eaon.dev/v1/chat/completions

{
  "model": "deepseek-v4-flash",
  "messages": [
    { "role": "user", "content": "Hello!" }
  ],
  "web_search": true
}

{
  "choices": [{ "message": { "role": "assistant", "content": "..." } }],
  "eaon": {
    "provider": "navy",
    "providerName": "Navy API",
    "model": "deepseek-v4-flash",
    "attempted": 1,
    "plan": "instant"
  }
}

Simple text

Pollinations-style plain-text endpoint — no JSON body required.

GEThttps://api.eaon.dev/v1/text/{your prompt here}

Optional model query parameter (defaults to auto).

curl "https://api.eaon.dev/v1/text/Write%20a%20haiku%20about%20coding"
curl "https://api.eaon.dev/v1/text/Write%20a%20haiku%20about%20coding?model=gemini-3.1-flash-lite"

Web search

Set "web_search": true on chat completions, or append :online to any model id (e.g. deepseek-v4-flash:online). Eaon fetches live results via MIKLIUM Search and injects them into the prompt.

GEThttps://api.eaon.dev/v1/search?q=latest+AI+news

{
  "search": ["latest AI news"]
}

Models

GEThttps://api.eaon.dev/v1/models

GEThttps://api.eaon.dev/v1/models/catalog

Click any model to copy its ID for use in the model field.

Full catalog with descriptions: browse all models.

Provider status

GEThttps://api.eaon.dev/v1/providers

GEThttps://api.eaon.dev/v1/uptime

/v1/uptime returns per-model uptime, latency and heartbeat history from the monitor's cached state — calling it never triggers a check. See it rendered.

End users never need provider keys. Eaon routes requests server-side using configured upstream keys. Live uptime: status dashboard.

Errors