Skip to content
KeenKind.Docs
Console

API quickstart

Use the KeenKind API in your existing coding tool. You can connect directly without installing the KeenKind desktop app. Follow these five steps.

1. Prepare your account and API key

Jump to 1. Prepare your account and API key
  1. Open the KeenKind console and sign in or register.
  2. Go to API keys and create a key for the tool you plan to use.
  3. Save the key after creation. The full key is shown only once. Keep it out of repositories and public screenshots.

If the create button is missing or unavailable, contact support to check your permissions. Model requests require available account balance or credit. Confirm purchase terms and current rates before use; see Models and billing.

2. Get the endpoint and model ID

Jump to 2. Get the endpoint and model ID
OpenAI Base URLResponses / Chat Completions
https://api.keenkind.ai/v1
Anthropic Base URLMessages
https://api.keenkind.ai
  • API key: the KeenKind key you just created.
  • Model ID: an id from your account’s current model catalog.

Tools use different protocols. Choose the endpoint in the corresponding guide. Get model IDs from the current model catalog instead of copying a display name from another platform.

3. Configure a tool you already use

Jump to 3. Configure a tool you already use

Choose your tool and follow its setup guide.

For your own programs, see API request examples.

4. Complete a small coding task

Jump to 4. Complete a small coding task

Open a familiar small project, select the service and model you configured, and send:

Read this project's README. Explain what the project does, how to start it, and how to run its tests.
Say when information is missing. Do not modify any files yet.

Check that the tool reads project files and that the answer matches the project. Then try a small fix or test before starting a long task.

Open usage records and check the request time, model, and charge. One coding task may trigger multiple model requests, so review the records for the task’s time range.

If setup fails, follow API troubleshooting. After an interrupted response or an uncertain result, check the records before retrying.