Your AI subscriptions.
One API.

Yes. With LLM To API, you can use your Codex subscription through an AI API. Connect your own Codex account, get a Base URL and API key, and use them in compatible AI clients, your own apps and automated workflows.

Supports Chat Completions and Responses using your existing subscription allowance.

Claude Code, Kimi Code and similar coding subscriptions can also be exposed through an AI API with an account adapter, bringing subscription usage to AI programming, vibe coding and your own tools.

Available now: Codex. Claude Code and Kimi Code subscription integrations are not yet available on this platform.

Use it for AI chat, coding assistance, text processing and AI features in your apps. Configure a client that accepts a custom API endpoint, or send requests from your own code and pass the results into a workflow. Available capabilities depend on the connected model and protocol.

Model availability depends on your provider account, connection checks and proxy configuration. After connecting an account, check the models available to that connection and use the model names shown in your configuration. Provider catalog counts do not mean your account has access to every listed model.

You need an email address, your own AI provider account and a client or app that accepts a custom API endpoint. Register and verify your email, connect your provider account, check the connection, then create a proxy to get a Base URL and API key.

Creating a live proxy requires an active platform plan. Current pricing and checkout availability are shown on the pricing and billing pages.

Yes. If your client has settings for a Base URL, API key and model name, enter the values from your workspace. You only need to write code when integrating AI features into your own program.

Yes, if it supports a custom API endpoint and a protocol available on your connection. Replace its Base URL and API key with your workspace values and select an available model. Clients that only use a fixed provider endpoint cannot connect directly.

The platform provides formats including Chat Completions, Responses and Anthropic Messages. Codex subscription connections support a Chat Completions compatible endpoint and Responses. Follow your workspace configuration examples when selecting a protocol and model; supported features vary by connection.

The hosted API endpoint does not require you to deploy a gateway or keep a personal computer running to forward requests. Your client or app needs an internet connection when making requests, and your provider authorization and platform plan must remain active.

No. The platform fee covers API access, hosting and management. You supply your own provider account. Subscription connections use that account’s available benefits; ordinary API-key connections follow the provider’s API billing. Budget for the platform plan and your provider subscription separately.

API access does not increase or reset your provider allowance. Requests still use the original account’s available usage and remain subject to its model access and rate limits. If the allowance runs out, wait for it to renew or use another supported connection with available usage. Repeated retries do not restore an allowance.

Yes. Manage multiple connections in one workspace and create proxies and API keys for different uses. Your plan sets limits for accounts, accounts per provider, proxies, keys and total concurrency. Provider limits also apply; adding more apps does not provide unlimited concurrency.

Yes. Create a key on the API keys page, choose its proxy, and optionally restrict models or set an expiry date. Separate keys make it easier to track their purpose and pause or revoke access for one app. The full key is displayed only when it is created, so save it at that point.

The platform uses your provider authorization or API credentials to connect upstream and encrypts stored credentials. Clients use the API keys generated by the platform, so you do not need to enter provider credentials into every client. Keep both types private and leave them out of support tickets, screenshots and public code.

Your workspace shows usage, request status, models, latency and errors. Request-body capture is controlled in workspace settings. When enabled, new request bodies are stored according to your plan and retention settings and can be deleted through the workspace. Bodies that were never captured cannot be recovered.

Model usage reflects the data returned by the provider. Requests without complete usage data are marked as unknown.

First check that the Base URL, API key, model name and protocol match your workspace configuration. Then check your plan, key expiry and connection status. If provider authorization has expired, follow the connection’s instructions to reauthorize or update credentials. For rate limits or exhausted allowances, reduce request frequency or wait for usage to become available.

If the issue continues, open a support ticket with the request ID, time and error message so the request can be traced.

Revoke the affected key on the API keys page, then create a new key and update the clients that should retain access. For a temporary stop, pause the key or proxy. A lost full key cannot be displayed again; create a replacement.

Yes. Choose Chinese or English when registering; the workspace uses that language after sign-in. You can also switch languages at the bottom of the sidebar, and the choice is saved to your account. Change your time zone in workspace settings to display request times and other records in your local time.