Know the request shape before the endpoint is live.

The intended surface is a chat completions endpoint. Arnict operates the model-serving path directly on US-based infrastructure; requests are not forwarded to a third-party inference provider.

Requests authenticate with a bearer API key created from your dashboard. API keys must remain server-side: never embed them in browser code, mobile builds, or public repositories.

Authorization: Bearer $ARNICT_API_KEY

Stored key records are hashed. The full secret is shown once at creation; afterwards only the prefix and metadata are displayed.

Use kimi-k3 as the model value. Roadmap models will receive their own identifiers when released.

{
  "model": "kimi-k3",
  "messages": [
    { "role": "system", "content": "You are a precise assistant." },
    { "role": "user", "content": "..." }
  ]
}

The response follows the familiar chat completion object, including usage fields that map to the published token rates for metering.

curl https://api.arnict.com/v1/chat/completions \
  -H "Authorization: Bearer $ARNICT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kimi-k3",
    "messages": [
      {
        "role": "user",
        "content": "Explain directly hosted inference."
      }
    ]
  }'

These examples are labelled illustrative until the inference endpoint is operational. The status page is the authority for current availability.

Pay As You Go includes $5 of free equivalent API usage monthly. Developer Pass ($29.99/mo) provides $40 per rolling week and $120 per billing month. See pricing for allowance rules.