> ## Documentation Index > Fetch the complete documentation index at: https://apione.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # Frequently asked questions > Answers about endpoints, models, clients, data, and charges. Verify the endpoint, key, and model with a minimal request before checking optional client features. | Symptom | First action | Details | | - | - | - | | 401, 404, or an unavailable model | Check the address, authentication, model ID, and key permissions | [Troubleshooting](https://apione.ai/docs/en/support/errors.md) | | cURL works but an SDK or client fails | Check the actual request URL, environment variables, and proxy settings | [SDK setup](https://apione.ai/docs/en/guides/sdks.md) | | HTTP 200 with empty or incomplete output | Inspect finish reasons, tool calls, and stream errors | [Streaming](https://apione.ai/docs/en/models/streaming.md) | | Unexpected charges | Check input, output, cache usage, and the key's billing group | [Usage and billing](https://apione.ai/docs/en/account/usage.md) | ## Should the Base URL include /v1? Use `https://api.apione.ai/v1` with the OpenAI SDK and `https://api.apione.ai` with the Anthropic SDK. If a client requires a complete endpoint, enter the full path. See [addresses and authentication](https://apione.ai/docs/en/api/overview.md). ## Can I use another provider's key? Requests to APIONE require a key created in the APIONE console. Keep each key paired with the correct provider's address. ## Why is an example model unavailable? An example model may not be enabled for your key. [List models available to the key](https://apione.ai/docs/en/models/catalog.md), copy the exact ID, and check permissions, billing group, and balance. A display name in the model directory is not necessarily the request ID. ## Does OpenAI compatibility include every feature? The API, model, and client must all support the feature. Check image input, structured output, tools, and Responses separately. See [model capabilities](https://apione.ai/docs/en/models/catalog.md) and [structured output and tools](https://apione.ai/docs/en/models/advanced.md). ## Why does a conversation forget earlier messages? For Chat Completions, your application sends the required message history. A new user message alone does not include earlier requests. See [chat requests](https://apione.ai/docs/en/models/chat.md). ## Why is an HTTP 200 response incomplete? Check `finish_reason`, stream errors, and disconnects. Tool calls may have no ordinary text; `length` can indicate truncation. HTTP 200 confirms the response or stream started successfully, so also inspect its final result. See [streaming](https://apione.ai/docs/en/models/streaming.md). ## The console shows the key is valid, but the API says it has expired `API token has expired` means the API returned an expiration error. If the console shows the key is still valid, you need to verify the authentication record further. 1. Confirm the client is using that key. Check for old environment variables, old processes, or duplicate authorization headers. 2. Call `GET https://api.apione.ai/v1/models` directly and keep the status code and response body. 3. Record the request time and time zone, as well as `X-Request-ID` or `X-Maxx-Trace-Id` in the response; keep any other request identifiers if returned. 4. Submit the console valid status and API error to support for investigation, focusing on validity storage, time units, permanent validity values, and auth caching. Use the [support template](https://apione.ai/docs/en/support/contact.md) when submitting issues. Provide only the page-recognizable name or redacted tail of the key. ## Are failed requests charged? Model calls finally classified as failed by the platform are not charged. A client timeout, cancellation, disconnected stream, or failed download after successful generation does not by itself mean the platform call failed. If generation completed or billable usage was confirmed, the applicable model and group rules determine the charge. For streaming, check the final status and actual usage rather than HTTP 200 alone, the client's error message, or whether the client received every output chunk. Incomplete delivery does not prove that no generation occurred. For a disputed status or charge, retain the request ID, timestamp and timezone, model, final status, and error details, then contact support. Confirmed erroneous or duplicate charges are corrected after investigation. Avoid repeatedly creating tasks while the original outcome is unknown. ## Can I put an API key in browser code? Keep the platform key on your server. Authenticate and limit end users at your application layer. Keep keys out of public browser code, logs, and screenshots. See [key management](https://apione.ai/docs/en/getting-started/api-keys.md). ## Is zero retention or a fixed service level guaranteed? See APIONE's current [privacy information](https://apione.ai/en/api-docs/docs/privacy) and [terms](https://apione.ai/en/api-docs/docs/terms) for data processing and service conditions. Confirm applicable conditions with [support](https://apione.ai/docs/en/support/contact.md) before handling sensitive data or arranging enterprise service. ## Is APIONE always priced the same as the official provider API? No. APIONE charges depend on the model price, the API key's billing group, and measured usage. Current groups can often be below the provider's official price, but the final charge is the current model, console, and billing record, not a permanent fixed discount. ## Do web chat and API calls share a balance? Yes. Web chat, images, video, and API usage share the APIONE balance. Credit priority and expiration depend on the current account and plan rules. ## Does provider failover guarantee success? No. Smart routing can choose another available route for the same model, but it cannot fix invalid parameters, permission, model, or balance errors. A request can still fail when every route is unavailable. ## How do spending points work? Web chat and API usage earn 5 points per \$1 spent. 100 points redeem \$1 in usage credits, equivalent to 5%. Points cannot be withdrawn and expire 6 months after they are credited. ## How are referral and first top-up rewards calculated? The inviter earns 5 points for every \$1 an invited friend tops up; those points are credited after a 7–15 day freeze. The invitee earns 5 extra points per \$1 on the first top-up, capped at 500 points (\$5 in usage credits) for that transaction. The same device, IP, or payment account qualifies once, and abnormal activity invalidates related rewards. See [Points and rewards](https://apione.ai/docs/en/account/rewards.md). ## Do plans renew automatically, and does upgrading restart validity? Plus, Pro, and Max are one-time purchases. All credits are issued at purchase and remain valid for 28 days; there is no daily issuance or automatic renewal. Upgrades add prorated credit differences and charge prorated price differences while preserving unused credits and the original expiry. Downgrades are not supported. At expiry, unused plan credits are cleared and you can buy another plan yourself. The separate pay-as-you-go balance is unaffected. See [usage and plan rules](https://apione.ai/docs/en/account/usage.md) for prices and formulas. Effective group prices already include the group discount; do not multiply that discount again. Apply the group rule once only to a price explicitly labeled as a base price. See [pricing and reconciliation](https://apione.ai/docs/en/account/pricing-groups.md). This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.