Use AI models through one API

Before you start

• Create an ApiClo account and a separate API key for your application.
• Check the current model list and tariff in your account before choosing a model; availability and prices can change.

Steps

1. Choose an interface

   Use the OpenAI-compatible /v1 endpoint for a supported client. Anthropic Messages clients use their documented endpoint settings instead; do not mix the two URL conventions.

2. Copy the exact model ID

   Select a model from the current catalog and copy its API identifier unchanged. A family name such as GPT or Claude is not itself a guaranteed model ID.

3. Make a small test request

   Use a protected API key, a short prompt and non-streaming text first. Check the response and the usage record before enabling streaming, tools or production traffic.

Check the result

Compare the model ID and tariff shown in your account with the request and its usage record. Test the exact interface and optional features your application needs; a text response alone does not prove full compatibility.

Troubleshooting

• 401: confirm the key is active and belongs to ApiClo; never paste it into public code.
• 404 or model error: recheck the Base URL, endpoint family and exact current model ID.
• Capacity or provider error: retry only according to your application's policy and inspect the documented fallback behavior before relying on it.

Compatibility and billing

API usage is billed by the selected model's current tariff. Promotional free aliases are intended for the ApiClo web chat when capacity is available, not for external API clients. ApiClo is an independent service, not an official OpenAI or Anthropic product; model availability and optional features are not guaranteed for every client.