> ## Documentation Index > Fetch the complete documentation index at: https://docs.beyondguard.ai/llms.txt > Use this file to discover all available pages before exploring further. # Token Management for Endpoint Access > Issue, scope, and rotate the access tokens that authenticate integrations to your BeyondGuard endpoints — with per-token rate limits and multiple tokens per endpoint. **Token Management** is the authentication layer for every integration and API access to the AI Proxy. From the **Token Configuration** screen you create tokens for specific purposes, describe them, and configure their access limits in detail. ## Creating tokens Tokens are generated automatically by the system and let you establish secure integrations with external services. For each token you can set: * **Description** — to clearly record its intended use. * **Expiration date** — a validity period aligned with your security policy. * **Rate limiting** — both a per-minute request limit (**Rate Limit per minute**) and a total usage quota (**Total Request Limit**), on a per-token basis. These limits reduce the risk of misuse, excessive queries, and unauthorized access, making token-based traffic controllable at enterprise scale. ## Multiple tokens per endpoint Each endpoint supports **multiple independent tokens**. You can define separate tokens for different integrations, applications, or environments — such as **dev, test, and prod** — and manage each token's usage, revocation, and renewal independently. This matters for safe rotation: | Benefit | What it means | | - | - | | **Safer token rotation** | Rotate or revoke a single token without interrupting every integration on the endpoint. | | **Reduced outage risk** | A compromised or expired token can be disabled while other clients keep operating. | | **Clearer access separation** | Keep environments and integrations isolated with dedicated tokens. | | **Stronger accountability** | Track which client accessed the system with which token, and when. | Previously, a single token per endpoint meant any rotation (leakage, policy, or integration change) could disrupt every client at once. Multiple-token support removes that single point of failure — prefer a dedicated token per integration and environment. Token management is the final authentication step before a request reaches the endpoint defined under [Policy Settings](/platform/policies). For programmatic use, tokens are passed as described in the [Guard Gateway API](/api-reference/guard-gateway) reference. ## Related How tokens authenticate API requests. The endpoints tokens grant access to. The policy set applied behind the token. Identity and access across the platform. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.