The Developers page

Developers is where you manage the pay-as-you-go API. Open it from the left bar, or go to /developers. The page itself is headed API. It works on every account, including Free.

At the top right, Documentation opens the API guide, and the round arrow refreshes every figure on the page.

The API wallet

API keys pay from a prepaid API wallet. It is separate from your plan: API keys never spend your plan’s credits, and your plan never spends the wallet.

The summary at the top shows your Available API balance, in dollars, next to what you Spent, how many Requests you made and your Tokens in / out over the last 30 days. While a call is running, part of the balance can show as reserved for calls in progress. That is the most the call could cost. When it finishes, you pay what it used and the rest comes back.

Add funds

1. Click Add funds.
2. Enter an amount in dollars, between $5 and $500. The dialog shows how many credits you’ll get, at 1,000 credits per dollar.
3. Click Pay with crypto. You’re taken to a checkout page to pay.

The credits are added once the payment is confirmed. API credits you buy never expire. Crypto is the only way to pay for API credits for now. If the page says “Buying API credits is not available on this account yet.”, contact us.

Tabs

Tab	What it shows
Overview	Your keys, a quick start, the last 30 days of usage and your most recent requests, on one screen
API keys	Create, limit and revoke API keys
Usage	Spend per day, by model and by key
Requests	The history of your calls, one row each
Prices	What each model costs per million tokens

API keys

To create a key, click Create key, give it a name so you can tell your keys apart, and click Create key again. The page shows “Copy your new key now. It will not be shown again.” Copy it, store it somewhere safe, and click I have saved it.

Each key in the list shows its name, the start of the key (ak_…), when it was created and last used, its monthly cap and its rate limit. The … menu next to it has two actions:

• Limits sets a Monthly cap in dollars and a number of Requests per minute. Leave the cap empty for no cap, and the rate empty for the default of 60 a minute. The highest rate you can set is 600 a minute. The cap counts what the key spent in the calendar month (UTC), including calls still running, and resets on the 1st. Once a call would take the key over its cap, it is refused with key_spend_cap_reached.
• Revoke asks you to confirm with Revoke key. Apps using that key lose access at once. You can’t undo it, but you can create a new key.

You can create a key before adding funds, but it can’t make calls until the wallet holds credits. An account can hold up to 20 active keys, terminal keys included.

Only API keys are listed here. Terminal keys, for Aderra Agent, Claude Code and Codex, are made on the Connect page and spend your plan’s credits.

Quick start

On the Overview tab, Make your first request shows a ready-to-run example with your base URL and the default model, in four tabs: cURL, Python, JavaScript and Anthropic format. Every example reads the key from the ADERRA_API_KEY environment variable, so you never paste the key into code. More examples, including streaming, are in the API guide.

Usage

Usage covers the last 30 days: Daily API spend as a chart, then the same spend By model and By key. Web searches show as their own line, Web search. Failed calls are counted but not charged; when there were any, the tab says how many.

Requests

Requests lists your calls, newest first. Load more shows older ones. Each row has the Time, the Key, the Model, the Endpoint (for example /v1/chat/completions), Tokens in / out, the Cost, the Duration and the Status: Done, Failed (not charged) or Running.

The history holds metadata only. It never shows what you sent or what the model answered.

Note

Usage and Requests include calls made with your terminal keys too. Those were paid from your plan’s credits, not the API wallet, and their cost is the plan credits they used, shown in dollars at 1,000 credits = $1.

Prices

Prices lists each model with its price per million tokens for Input, Cached input and Output, paid from the API wallet. Cached input is the part of your request the model has seen recently, and costs less.

Most models cost the same through the API as on a plan. The unfiltered model, GLM-5.3 Flash Unfiltered, costs more through the API. Its row also shows “On a plan: … in / … out” with a See plans link, for comparison.