Pricing and billing

Pay as you go

Your balance is in Rupiah and goes down automatically as each request finishes. There is no monthly fee. All model prices are on the pricing page, per 1 million tokens.

The cost of one request:

cost = (input tokens x input price)
     + (output tokens x output price)
     + (cache read tokens x cache read price)
     + (cache write tokens x cache write price)

All prices are per 1 million tokens. The exact token counts for each request are in usage in the response and in the Usage menu of the dashboard.

Cache pricing

Coding agents resend the same prompt beginning at every step: the system prompt, the tool list and the history. The part that matches an earlier request is read from cache and billed at the cache read price, which is much lower than the normal input price. The Cache read column on the pricing page shows it per model.

To keep caching effective:

• Do not change the start of the prompt on every request, for example by putting the time or a random number in the system prompt.
• For Claude Code, set CLAUDE_CODE_ATTRIBUTION_HEADER=0 (see Claude Code).

Balance check

Before a request runs, Vikey estimates its maximum cost from the input tokens and the output limit (max_tokens). If your balance is not enough, the request is rejected with 402 and is not charged.

Send a sensible max_tokens. A very large output limit makes the estimate large too, so a request can be rejected even though your balance would cover the actual answer.

Extra charges

Part	Cost
Vikey Auto model selection	Rp10 per new message, only for Vikey Auto model ids
Image models	Per image, listed on the pricing page

There are no other charges beyond what is recorded in Usage and the transaction history.

Usage and billing

Dashboard menu	Contents
Overview	Today’s usage and recent transactions
Usage	Per model: requests, tokens and cost
Billing	Balance, top ups, transaction history, invoices

Detailed per-request logs are kept for 7 days. The daily totals that billing is based on are kept permanently.