Models and billing

Confirm the model ID and required capabilities before configuring a tool. See Model pricing for current and original prices in USD. For console instructions, see Usage and billing details and Balance, top-ups, and orders.

Get current model IDs

Use this endpoint to list the models available to your account:

GET https://api.keenkind.ai/v1/models

Set an environment variable first, replacing the placeholder with your API key.

macOS / Linux

Bash or Zsh:

export KEENKIND_API_KEY='YOUR_KEENKIND_API_KEY'
curl --silent --show-error 'https://api.keenkind.ai/v1/models' \
  --header "Authorization: Bearer $KEENKIND_API_KEY"

Windows PowerShell

$env:KEENKIND_API_KEY = 'YOUR_KEENKIND_API_KEY'
$headers = @{ Authorization = "Bearer $env:KEENKIND_API_KEY" }
Invoke-RestMethod -Uri 'https://api.keenkind.ai/v1/models' -Headers $headers |
  ConvertTo-Json -Depth 10

Copy an id from the returned entries into your tool’s or request’s model field. Listing models does not start a model generation request.

The catalog depends on account permissions and model and service status. If it is empty or a required model is missing, contact support instead of repeatedly guessing model names.

Choose an API

Purpose	Base URL	Full request path
OpenAI Responses	https://api.keenkind.ai/v1	/v1/responses
OpenAI Chat Completions	https://api.keenkind.ai/v1	/v1/chat/completions
Anthropic Messages	https://api.keenkind.ai	/v1/messages

Use the base URL in a tool or SDK configuration; use the full path for a direct HTTP request. Confirm tool calling, images, context length, and reasoning parameters for each model. The API name alone does not establish support.

Check balance and usage

Task	Console entry
Review model requests and charges	Usage records
Check balance and top-up options	Balance and top-ups
Review payment orders	Orders
View or manage key limits	API keys

Account balance and key limits are separate checks. A key can stop working after reaching its daily, monthly, or fixed limit even when the account still has balance.

Check the cost of a coding task

One task can involve multiple model requests. Note the start and end time, find the requests in that time range, and total their charges. Use separate keys for different tools to distinguish the source.

Check model IDs, input and output usage, and any separately billed cache reads or writes. See Model pricing for prices. Historical requests retain their price snapshots, and a tool’s estimates may differ from console settlement. A real coding-task cost example will be added after usage has been collected.

An interrupted connection does not prove that no request ran or no fee was incurred. Requests with partial output may be billed for confirmed usage. Check records before resending; see API troubleshooting.