Skip to content
Documentation menu

API referenceDraft

Credits

How API usage is paid for.

Draft

The hosted API is a planned service with a draft specification. The local API described alongside it already exists and runs on your own machine.
  • Credits are a unit of consumption, not dollars.
  • Each use costs credits by function: text, search, vision or deep research.
  • Real token counts are recorded separately from credits.
  • Every change to a balance is recorded as its own transaction.

On every plan

  • Paid plans are billed monthly with PayPal or card.
  • Queue priority (P1, P5, P6, P7) is a plan's relative place in the request queue. It says nothing about model quality, intelligence or capability.
  • Request allowances and rate limit values are not published yet. They will be set once KOLDOS has real capacity, load and usage data.
  • Rate limits apply per IP address, account, API key, endpoint and device.
  • Responses stream as server-sent events.
  • A daily budget stops spending before it surprises anyone.

Cost per use

  • Text: 1 credit per use.
  • Search: 2 credits per use.
  • Vision: 3 credits per use.
  • Deep Research: 8 credits per use.

Paying for a plan

Credits come with the plan. Paid plans are billed monthly with PayPal or card. The hosted API has a planned launch window from 2026-12-30 to 2027-02-10, not a fixed date.

Checking your balance

The dashboard shows your balance, every grant and daily usage.

Running out

A request that would exceed your balance returns 402 insufficient_credits.