Connect your API client to ApiClo

Before you start

• Create an ApiClo account and a separate API key for this integration.
• Check your balance and the current model list in your account. Copy the exact Model ID; a display name is not an API identifier.

Steps

1. Set the endpoint

   Enter https://apiclo.com/v1 as Base URL. A client that appends /chat/completions then targets /v1/chat/completions. Do not append /v1 twice.

2. Add authentication

   Store your ApiClo key in the client's protected key field or a backend environment variable. For direct HTTP, use Authorization: Bearer YOUR_APICLO_API_KEY. Never commit the key or include it in public HTML.

3. Select a model

   Copy a Model ID from the current ApiClo catalog and enter it unchanged. Start with a short text request and stream=false before testing streaming or tool calls.

Check the result

Send one short test message from your own client. Check the HTTP status, returned content and usage entry in ApiClo. A successful text response does not establish compatibility with every optional API feature.

Troubleshooting

• 401: check that the key belongs to ApiClo, is active and has no extra spaces. Do not share it in support messages.
• 404: check the final URL, especially a duplicated /v1 or the wrong endpoint family.
• A model error: copy the current Model ID again and check that this model is available for the selected interface.
• A streaming or tool error: first repeat with a minimal non-streaming text request; then check the client-specific guide.

Compatibility and billing

OpenAI-compatible and Anthropic Messages clients use different endpoint conventions. Clients that append /v1/messages themselves use https://apiclo.com without /v1. The selected model is tried first; documented provider or capacity failures can trigger controlled fallback at the selected-model tariff. Managed Hermes hosting is charged separately. ApiClo is an independent service, not an official OpenAI or Anthropic product.