
## What the tunnel stores {#storage}

The **Save request** and **Save response** settings apply independently to new requests. A disabled direction has no saved HTTP data. When storage is enabled, the request and response can include headers, trailers and bodies. The journal still records the HTTP outcome and byte counts independently of these settings. Changing storage settings does not recover data for earlier requests.

HTTP details and bodies expire 90 days after receipt. Check **Delete after** in the request card. The journal entry can remain after the HTTP data expires. See [journal and statuses](/en/docs/tunnels/history).

## Inspect the data {#inspect}

Open [Tunnels → Requests](https://dashboard.adal.cloud/tunnels/requests), select a request and look at **Request** and **Response**. Repeated headers have separate rows. Trailers and informational responses appear when saved. Use **Download bytes** on an escaped field value to retrieve its original bytes.

The body cards show size and storage state. JSON, text, form data and supported images can be previewed up to 256 KiB. Use download for larger bodies or compressed content.

## Download the saved bytes {#download}

Choose **Download request body** or **Download response body**. The browser downloads the original stored bytes up to 128 MiB, without decompressing them. Use **Cancel download** to stop. If a body was only partly saved, the file contains the saved prefix. Check the completion message before using it as a full copy.

| State | What you can retrieve |
|---|---|
| **Saved** | The saved body. A saved body may be empty. |
| **Saving**, **Partially saved** | A saved prefix when bytes are available. Check whether the download is complete. |
| **Storage disabled**, **Storage failed**, **Data expired** | No body is available for this direction. |

Saving data and billing are independent. [Tunnel usage](/en/docs/common/traffic#tunnel) is calculated from the body bytes actually received even when storage is disabled. Treat downloaded data as private. Headers, URLs and bodies can contain credentials and personal information.
