Connect ZCode

For ZCode versions with custom providers. Entry points follow the official model settings guide; interface names may change between versions.

1. Open model settings

Click the model name in the chat composer to open the selector. Click Manage models at the bottom to open model settings.

2. Add KeenKind

At the bottom of the provider list, click Add provider and enter:

Field	Value
Name	KeenKind
OpenAI endpoint	https://api.keenkind.ai/v1
API key	Your KeenKind API key

Use the /v1 URL for an OpenAI-compatible API. If a model is confirmed to require an Anthropic-compatible API, set the Anthropic endpoint to https://api.keenkind.ai instead. Verify one confirmed protocol first.

Use a custom provider. Do not enter your KeenKind key into a BigModel or Z.ai account subscription configuration.

3. Add and select a model

1. Click Add model and enter an actual model ID from the catalog.
2. Save and enable both the provider and model.
3. Return to chat and select the model under KeenKind.

Keep advanced parameters at their defaults initially. Context window, maximum output, and thinking parameters must match actual service support.

4. Complete a small task

Open a project and ask the model to read its README and explain startup instructions. Check file access and the answer, then review the call in usage records.

Symptom	Check first
Model missing from selector	Provider and model are saved and enabled.
401 / 403	KeenKind key and account permissions.
404	Full request path, API type, and model ID.
Thinking parameter error or failed tools	Restore default parameters, then check model and protocol capabilities.

See API troubleshooting for more help.

Based on ZCode’s model configuration documentation. Sources checked: 2026-10-18.