21. FAQ#

Q: How much do I need to change in my existing OpenAI code? A: Usually just two places — baseURL and apiKey; every other request parameter stays the same.

Q: Can I use the official OpenAI SDK directly? A: Yes. OpenAI SDKs in all mainstream languages support a custom baseURL — just point it at http://183.240.79.21:52099/v1.

Q: How do I call non-OpenAI models such as Claude / Gemini? A: We recommend routing everything through /v1/chat/completions and simply setting model to the model name — no protocol switch needed. If you must use the native Anthropic protocol, /v1/messages works (it does not support function calling / multimodal).

Q: How do I know which models my current key can use? A: Call GET /v1/models; whatever it returns is exactly the list of models you may access.

Q: My calls return 404 — what should I do? A: First confirm the path carries a single /v1 (e.g. /v1/chat/completions, not /v1/v1/...); then confirm the model name is spelled correctly and appears in the GET /v1/models response.

Q: How do I know when a stream has ended? A: SSE streams from OpenAI endpoints end with the data: [DONE] marker; the /v1/messages endpoint returns an Anthropic-shaped SSE event sequence.

Q: How long do video / image generation tasks take? A: They are asynchronous — poll /v1/tasks/{id} with the id returned at submission until the status becomes complete.

Q: I forgot to save my API Key plaintext — what now? A: For security, the plaintext is displayed only once at creation and cannot be retrieved. Revoke the old key in the console and create a new one.

The domain http://183.240.79.21:52099 in this document is a placeholder; replace it with the actual service domain once the service launches. The local development Base URL is http://183.240.79.21:52099/v1.