Static IP for IP Allowlisting: One Address for Remote Teams
Contents

- What does an IP allowlist check?
- Why do home and mobile connections fail an allowlist?
- Office IP, VPN or a static ISP proxy: which fits a team without an office?
- How do you set up one address for the team?
- What breaks an allowlist, and how do you keep the address?
- Why must an IP allowlist never be the only control?
- When is this the wrong setup?
- Quick answers
In short: An IP allowlist admits a connection by the address it comes from. Home broadband addresses can change, and mobile addresses are shared, so a team without an office needs one fixed address of its own. A static ISP proxy is one: every member connects with its username and password, and the systems you allowlist see a single address. Keep it next to per-person sign-in with two-factor authentication.
An IP allowlist lets a system accept connections only from the addresses you list. It is easy to set up when everyone works from one office and awkward when nobody does. This guide is for a distributed team that has to reach systems it runs or is authorised to use, such as a code host, an admin panel or a cloud server, from one fixed address. ProxyJam sells ISP proxies, one way to get that address, and the guide also covers the two other ways and where each one fits. Our editorial policy explains how articles like this are sourced and corrected.
What does an IP allowlist check?
The source address of the connection that reaches it, and nothing about the person behind it. Each of the three common places a team meets one works the same way.
- A code host. GitHub’s IP allow list decides “whether users or apps can access protected resources through the web UI, APIs, or Git”. It takes a single address or a range in CIDR notation, and it “applies to users with any role or access, including enterprise and organization owners”. It is available to organisations on GitHub Enterprise Cloud.
- Office software. Google Workspace’s Context-Aware Access can build an access level from an IP condition, which “specifies an IP address range from which a user can connect to an app”, and from the countries a user connects from.
- A cloud server. An AWS security group rule names the source it admits, and a single IPv4 address is written with the /32 prefix length, as in 203.0.113.1/32.
When a connection goes through a proxy, the system sees the proxy’s address. One address can therefore stand for a whole team, and the allowlist cannot tell the members apart.
Why do home and mobile connections fail an allowlist?
They do not hold one address that belongs to one subscriber. Home broadband often gets its address by what the DHCP standard, RFC 2131, calls dynamic allocation: an address assigned “for a limited period of time”. When it changes is up to the provider. An entry that matched a member’s address on Monday can stop matching on Thursday, and the member is locked out until someone with admin rights updates the list.
The workaround GitHub’s documentation offers for this is a CIDR range wide enough to “include dynamically assigned addresses”. That keeps the member in, and it also lets in every other subscriber of that provider whose address falls in the range.
Mobile connections have the opposite problem. Carriers commonly put many subscribers behind one public address through carrier-grade NAT (RFC 6888), which our comparison of mobile and residential proxies walks through. The IETF’s review of address sharing, RFC 6269, states in section 13.2: “Simple address-based identification mechanisms that are used to populate access control lists will fail when an IP address is no longer sufficient to identify a particular subscriber.” An allowlist entry for a phone’s hotspot can admit strangers. Hotel and co-working networks add a third case, with an address nobody on the team controls.
A team without an office therefore needs an address with two properties: it stays the same, and nobody outside the team sends traffic from it.
Office IP, VPN or a static ISP proxy: which fits a team without an office?
Any of the three can give a team one fixed address. They differ in what goes through it and in what someone has to keep running, as the table Office line, VPN or static proxy sets out.
An office line
A business internet plan usually comes with a static address, and GitHub’s documentation uses exactly this case as its example: allowing access “exclusively from the IP address of your office network”. It works for the people in the building. Remote members then need a way onto the office network, which in practice is a VPN into it.
A VPN with a fixed exit
A VPN sends a device’s traffic, or the part a split tunnel selects, out through a server, and the allowlist sees that server’s address. You can run the server yourself on a cloud machine with a fixed public address, or buy a business VPN plan that includes a dedicated one. It is the choice when desktop apps, SSH sessions or anything without a proxy setting must reach the allowlisted systems. The cost is a client on every device and someone who runs the server and its accounts.
A static ISP proxy
A proxy carries only the traffic you send to it: a browser through an extension, a script or a tool through its proxy setting. For a team whose allowlisted systems are web apps, admin panels and APIs, that is usually enough, and there is no server to run. Our addresses are on a named internet provider’s network, served from data-centre hardware, as the definition of ISP proxies explains, and each one stays with you for the period you pay for.
For the allowlist itself, the network type makes no difference. A cloud server’s datacenter address passes an allowlist exactly as an ISP address does, and residential against datacenter proxies explains where the difference does count. If a VPN server on a cloud machine is something your team already runs, it may be all you need.
How do you set up one address for the team?
Order one address, give every member the same connection details, route the allowlisted systems through it, and only then add it to the allowlist. The steps are listed under Give the team one address. Three choices in them are worth explaining.
Authenticate with the username and password. Every address comes with its own host, port, username and password. An order can instead admit clients by an allowlist of their IP addresses, set through the API, and while that list is set the password stops working (integrations). For people working from home that is the same problem moved one step back: their changing addresses would lock them out of the proxy. Keep the password for people, and use the order’s allowlist only for a server or a test runner that has a fixed address of its own.
Route only the allowlisted systems. In the browser, ProxyJam Connector is a free extension for Chrome, Edge and Firefox. Its Auto mode lets a rule send just the allowlisted system’s hostnames through the proxy while everything else leaves from the member’s own connection, the approach covered in using a proxy only for certain sites. This also keeps the proxy within its session cap. Each address carries up to 100 concurrent sessions, whatever the traffic (Terms of Service, section 2.4), and a browser opens several connections at a time.
Stay on HTTP in Chrome. The order starts on HTTP, which carries HTTPS as a tunnel. Chrome cannot send a username and password to a SOCKS5 proxy, a limit explained in the proxy sign-in prompt, so a team on Chrome keeps the proxy on HTTP. Switching the protocol also gives the order a new port and new credentials, so settle it before you hand the details out.
Scripts, CI jobs and other code that call an allowlisted API take the same four values. The integrations guides show where Python requests, Selenium and Playwright expect them.
Before you switch enforcement on, keep a way back in. GitHub warns that losing access to the addresses on the list “could result in unintended consequences, such as getting locked out of your enterprise or organization”, and recommends more than one owner and a static network kept on the list for backup access. The list under Check these before the allowlist goes live covers the rest.
What breaks an allowlist, and how do you keep the address?
An allowlist breaks when the address it names stops being the one your team sends from. With a proxy, that happens in three ways.
- The period ends. The address is yours for the period you paid for, and renewing keeps the same address, as our ISP proxies page states. Turn on Auto-extend for it in My Proxies, which renews it from your balance, or put the renewal date in someone’s calendar. An address that is allowed to lapse takes the allowlist entry with it.
- You move to a different address. The address is not replaced during the period you paid for; a different one takes a new order and is assigned at random (Terms of Service, section 2.5). Add the new address before you remove the old one.
- A location rule disagrees. Workspace’s access levels can also check the country a user connects from. That country comes from a geolocation database, and databases can place one address differently. If yours moves the address out of its country, the location rule can refuse it while the IP entry still matches.
Changes on the system’s side take time too. GitHub notes that, because of caching, adding or removing an address “can take a few minutes to fully take effect”, so do the swap while someone is watching.
Addresses are priced per IP per period, on the pricing page, and a longer period costs the same per month. What it buys is a renewal date further away.
Why must an IP allowlist never be the only control?
An allowlist says where a connection came from, and nothing about who made it. Anyone holding the proxy’s username and password sends from the team’s address. NIST’s digital identity guideline, SP 800-63B-4, allows signals such as “authentication from an unexpected geolocation or IP address block” to lower the risk of a wrong sign-in, and states in section 2 that using them “does not impact or change the AAL of a transaction or substitute for an authentication factor”.
In practice that means three habits.
- One account per person. Every member signs in to each allowlisted system as themselves, with two-factor authentication turned on. GitHub, Workspace and AWS all check the address on top of the sign-in.
- Removal goes through accounts. When someone leaves, removing their account is what cuts their access. The address stays the team’s, and it would still admit a former member who kept the credentials.
- The proxy credentials are a shared secret. Keep them in the team’s password manager, give them only to people who use the allowlisted systems, and leave them out of chat messages and code repositories.
When is this the wrong setup?
The address is for systems your organisation owns or is authorised to use. Section 1 of our Acceptable Use Policy permits controlled network access for legitimate business purposes as long as it is not used to “act against a system’s access rules”, and section 5 prohibits accessing services in breach of their regional availability or platform access rules. Getting onto another organisation’s allowlist without its agreement is outside both. Three other cases call for something else.
- Regulated systems. If the allowlisted system belongs to financial services, healthcare or government, section 2 of the policy asks for verification before we open access to it.
- People outside your organisation. Section 6 forbids sharing or redistributing access to the service without our written permission, so a client or partner company needs its own arrangement.
- Traffic outside the browser and your code. Desktop apps, SSH and anything without a proxy setting need a VPN, or an office line.
For a team whose allowlisted systems live in a browser or behind an API, one static address is where it starts. The ISP proxies page lists the catalogue, 76 countries with a named carrier behind every address, and the pricing page shows what one address costs for each period.
