Why your mobile proxy IP changes
Understand carrier reassignment, modem reconnects and requested rotation. Learn which connection details to record when a mobile exit IP changes.
The proxy endpoint and the public exit IP are different. Your application can keep the same connection settings while the carrier assigns a different public address to the mobile connection.
Separate the endpoint from the exit IP
The host, port and credentials tell your client how to reach the proxy. A destination website sees the public address used by the carrier connection. That address is not necessarily a unique address reserved for your modem: carrier-grade NAT can share one public address across subscribers.
A dedicated modem or port therefore does not mean a permanently fixed public IP. Keep both the connection details and the observed exit IP in your troubleshooting notes, but redact credentials from any shared report.
Identify what happened before the change
| Event | What it may change | What to record |
|---|---|---|
| Requested rotation | The modem reconnects and may receive a different public IP. | Request time, response and exit IP after reconnecting. |
| Carrier reconnect or network event | The carrier may reassign the address without an application request. | Last successful request, interruption and first new IP. |
| New rental or allocation | The issued endpoint or credentials may also change. | Order reference and the connection settings currently shown in the portal. |
| Different client or network path | A request may bypass the proxy or use another configured endpoint. | Client settings, bypass rules and a same-client IP check. |
A requested rotation is not a promise of a previously unseen IP. A carrier can return an address that was used earlier.
Check the IP from the client that matters
Run a small IP-echo request from the application or browser context that performs your work. Comparing a proxied browser with an unconfigured command-line client can produce two different addresses without a carrier change.
Record UTC timestamps and a short sequence of observations around a reconnect. Keep the sample small and bounded. A monitoring script should stop or back off when the connection is unavailable rather than repeatedly issuing rotations.
curl --fail --silent --show-error \
--connect-timeout 10 --max-time 20 \
--proxy "$MP_PROXY_URL" https://api.ipify.orgHandle a change at a workflow boundary
Do not automatically replay a request that may already have changed server-side state. A timeout can occur after the server received the request. Use application-specific idempotency and recovery where available.
If your application requires the same public address for a fixed period, describe that requirement before buying and test it. Ordinary mobile connectivity cannot establish a guarantee about future carrier reassignment.
- Pause new work that depends on the existing session.
- Wait for connectivity to return using bounded retries.
- Confirm the exit IP and the intended destination behavior.
- Decide whether the application can continue the existing session or needs a new one.
- Resume only the work whose state you can account for.