Skip to content
KOLDOS
Sign in
KOLDOS
Documentation menu
/

Examples and SDKs

Getting startedDraft

KOLDOS API overview

The local API that exists today, the hosted API that is planned, and how they differ.

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.

There are two KOLDOS APIs. The local API runs on your own computer and exists today. The hosted API, KOLDOS as a service, is planned: the design is written and the service has not started.

Both speak the same common chat completions format, and in both a request is a whole KOLDOS turn: the Decision Engine picks a route, the Context Broker assembles the prompt, tools run and the result is verified. It is not a single call to the model.

Status

APIStatusWhere it runs
Local APIAvailable with KOLDOSYour computer, on 127.0.0.1
Hosted APIPlannedHosted GPUs. Not started

Planned hosted endpoints

MethodPathPurpose
POST/v1/chat/completionsA KOLDOS turn, not a single completion.
POST/v1/visionImage analysis.
POST/v1/searchWeb search.
POST/v1/researchDeeper research across several sources.
GET/v1/modelsModels available to your key.
GET/v1/creditsCredit balance and transactions.
GET/v1/usageConsumption per period.

What you need

  • For the local API: KOLDOS running on your computer. See Local API.
  • For the hosted API, once it exists: approved developer access. See Requesting access.
  • A Discord account to sign in to the dashboard.
  • An API key created in the dashboard.
  • Credits on your account.

Local or API?

ComparedLocal APIHosted API
Where the turn runsYour computerHosted GPUs
Needs a capable GPUYes, an NVIDIA GPU with CUDANo
Needs an internet connectionNoYes, for every request
AuthenticationOptional access tokenAPI key
CostYour own hardwareCredits, on plans from $0 to $14.99 a month
Your code and promptsStay on your computerSent to the service