# Authentication

For Genesis OpenAI-compatible APIs, prepare an access credential and send it with every request. This page applies to `GET /models`, Chat Completions, Embeddings, Rerank, and Responses. Follow the Messages API page for Anthropic Messages authentication.

1. Create a personal access token in the Genesis console, or create a service account with Genesis permission and the required model scope.
2. Store the token or API key in a local environment variable or secret manager.
3. Send `Authorization: Bearer <ACCESS_TOKEN>`.
4. List available models to confirm that the credential can call the target model.

```bash
export GENESIS_BASE_URL='<Genesis Base URL copied from the console>'
export GENESIS_ACCESS_TOKEN='<your-access-token>'

curl "$GENESIS_BASE_URL/models" \
  -H "Authorization: Bearer $GENESIS_ACCESS_TOKEN"
```

Use the complete base URL copied from the console. If it already ends in `/v1`, do not append another `/v1`. Do not store credentials in source code, images, logs, tickets, or screenshots.

## Choose a credential

| Credential | Use it for |
| --- | --- |
| Personal access token | Local development, command-line checks, or scripts running as you. |
| Service-account API key | Applications, CI/CD, and production automation. |

## Next steps

- [List available models](available-models.md)
- [Send a first Chat Completions request](quickstart.md)