API credits, spend limits and alerts

This page gets your SynapseX API workspace funded and keeps it from surprising you. When you finish you will have credits on the account, a monthly spend cap you chose, alerts that warn you before you reach it, and a clear understanding of the one error that stops everything dead.

How does billing work on the SynapseX API platform?

Pay-as-you-go with prepaid credits only. There is no monthly subscription and no seat licence — you buy credits, and requests consume them.

1 credit = US$1. (This is not the same unit as the credits inside SynapseX Chat, which are a completely separate balance on a separate product. See Accounts, plans and billing.)

What credit packs are available?

Scale and Growth include $25 and $100 of bonus credits respectively.

How do I buy credits?

1. Go to /dashboard/credits in the console.
2. Pick a pack.
3. Pay inline with the payment form that opens on the page. You are never sent to an external hosted checkout.

Two details that matter in practice:

• EU/EEA customers are charged in euros, with card, SEPA Direct Debit, iDEAL and Bancontact available. Everyone else is charged in US dollars with a card. The euro price of a pack is the same number as its dollar price — the Starter pack is $25 or €25 — and the credits granted are USD-denominated either way.
• A double-click cannot double-charge you. The purchase is idempotent, so an impatient second click reuses the same payment rather than starting a new one.

Credits are granted to the workspace once the payment succeeds.

What stops me running up a huge bill?

The hard stop is the balance itself. A balance at or below zero returns HTTP 402 on every API call, before any model runs:

{
  "error": {
    "message": "Insufficient credits",
    "type": "insufficient_quota"
  }
}

Because the check happens before the request reaches a model, you cannot run up a debt. The worst case is that your application stops working — which is exactly why you should set a spend limit and watch the alerts rather than relying on the 402 as your first warning.

How do I set a monthly spend limit?

On the Billing page (/dashboard/credits) set a monthly spend limit for the workspace. It is a US-dollar figure you choose, and you can change it at any time.

Two alert thresholds are configured by default: 50% and 80% of that limit. When you cross one, a badge appears on the Billing page telling you what percentage of the monthly limit you have used and which threshold was crossed. At 100% the badge changes to say the limit has been reached.

Where are my invoices and payment methods?

In the Stripe billing portal, which you reach from a button on the Billing page. That is where you download invoices and change the card on file — it is not part of the SynapseX console itself.

Where do I watch consumption?

/dashboard/usage. Choose a window between 1 and 90 days (the default is 30) and you get:

• total cost in US dollars, total tokens and total request count for the window;
• a breakdown by model, so you can see which model is actually consuming the budget;
• the most recent requests, each with its model, provider, token count, cost, latency and status.

A practical budgeting recipe

If you do not yet know what your workload costs, do not guess — measure it cheaply:

1. Buy the Starter pack ($25). It is enough to establish a real cost per unit of work.
2. Set the monthly spend limit below the pack, for example $20. You then get the 50% alert at $10 and the 80% alert at $16, well before the money runs out.
3. Run one representative day of traffic, then open /dashboard/usage with a 1-day window and read the total cost and the per-model breakdown.
4. Multiply by your real monthly volume, add headroom, and size the next top-up from that number instead of from a guess. Move the expensive share of traffic to a cheaper model first if the breakdown suggests it.

Related

• Manage workspace API keys — per-key spend limits
• Workspaces and teammates — who shares this balance
• Platform API errors