> ## Documentation Index
> Fetch the complete documentation index at: https://tyk.io/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Studio Administrator

> Overview of the AI Studio Administrator persona in Tyk AI Studio

The **AI Studio Administrator** is the primary manager of the Tyk AI Studio installation. This persona is responsible for the operational integrity, security, and financial governance of the AI platform. They ensure that AI resources are available, secure, and used within budget.

## Lifecycle

The typical lifecycle of an AI Studio Administrator involves four key stages:

1. **Configure**: Setting up the fundamental infrastructure, including connecting to [LLM Providers](https://tyk.io/docs/ai-management/ai-studio/llm-management.md) (e.g., OpenAI, Anthropic) and configuring [Vector Databases](https://tyk.io/docs/ai-management/ai-studio/datasources-rag.md) for RAG.
2. **Govern**: Establishing global guardrails, such as [Filters](https://tyk.io/docs/ai-management/ai-studio/filters.md) for PII masking, and [role-based access control (RBAC)](https://tyk.io/docs/ai-management/ai-studio/rbac.md).
3. **Onboard**: Creating [Teams](https://tyk.io/docs/ai-management/ai-studio/teams.md) and [Users](https://tyk.io/docs/ai-management/ai-studio/user-management.md), assigning roles, and allocating resource quotas.
4. **Monitor**: Continuously tracking organization-wide token usage, enforcing [Budgets](https://tyk.io/docs/ai-management/ai-studio/budgeting.md), and auditing security alerts via the [Dashboard](https://tyk.io/docs/ai-management/ai-studio/analytics.md).

## Core Features

### Provider Management

Centralized management of LLM credentials and configurations. Admins can add, update, or deprecate models from providers like OpenAI, Anthropic, or Mistral without requiring changes to consumer applications.

* **Key Capability**: Unified API key management (stored as [Secrets](https://tyk.io/docs/ai-management/ai-studio/secrets.md)) to prevent credential sprawl.

### Policy & Security

Enforcement of global security and compliance rules.

* **PII Detection**: Automatically redact sensitive information (credit cards, emails) from prompts before they reach the LLM.
* **RBAC**: Granular control over which teams can access specific models or tools.
* **Rate Limiting**: Protect upstream provider quotas with a rate limiting [plugin](https://tyk.io/docs/ai-management/ai-studio/plugins/overview.md). The Edge Gateway has no built-in rate limit settings.

### Cost Management

Comprehensive financial controls to prevent "bill shock."

* **Budgets**: Set monthly spending limits at the Organization, Team, or App level.
* **Alerts**: Receive notifications when spending approaches defined thresholds (e.g., 80% of budget).
* **Chargeback Reporting**: Detailed analytics to attribute costs to specific teams or projects.

### Audit Logs

Full traceability of system access and configuration changes.

* **Access Logs**: Track who accessed which model and when.
* **Configuration History**: In the Enterprise Edition, the audit trail records every change through the management API. It also records every login and SSO event. Refer to [Audit Trail](https://tyk.io/docs/ai-management/ai-studio/audit-trail.md).