OpenAI

What it enables in Codexa

• Store your own OpenAI API key encrypted in the workspace and test it from workspace settings before saving.
• Prepare the workspace for bring-your-own-key AI runs with OpenAI models (billed to your OpenAI account).

• test
• byok

Setup

Fields on the workspace connection form. Password-type fields are encrypted before storage and never shown again.

1. Open Settings → Integrations in your workspace (requires the admin or owner role).
2. Choose the provider, give the connection a name, and fill in the fields listed above. Secret fields are encrypted before they are stored and are never sent back to the browser.
3. Click Test connection. Codexa makes a single read-only request to the provider to confirm the credentials work.
4. Save. The connection is now available to projects in this workspace.
5. Today the AI agent runs with the operator-level OPENAI_API_KEY or ANTHROPIC_API_KEY; routing runs through a workspace key is being wired up and is not active yet. When no key is configured at all, the AI panel shows a 'not configured' state.

Required permissions and scopes

• An OpenAI API key with access to the chat completions API. Keys are encrypted at rest and never returned to the browser.

Limitations

• Usage limits from your Codexa plan (requests and tokens) apply to every AI run, including runs with your own key once that routing is active.
• Fine-tuned models and assistants/threads APIs are not supported.