On November 5, 2026, Anthropic added an IP allowlist to the Claude Console in beta. An admin or owner lists up to 1,000 public IPv4 or IPv6 addresses or CIDR ranges, and covered requests from any other address are refused (the Messages API returns a 403 permission_error), which limits where a leaked API key can be used.
Last updated: November 8, 2026
Table of contents
- What changed on November 5
- What the allowlist covers
- How to turn on the allowlist
- Entry rules and limits
- Avoiding and fixing a lockout
- Related beta: sign a member out everywhere
- Frequently asked questions
- Sources
What changed on November 5, 2026
The Claude Platform release notes for November 5, 2026 list two admin security additions. First, Claude Console organizations can set up an IP allowlist, in beta. Requests to covered services from addresses not on the list are refused, which limits where a leaked API key can be used. Second, Claude Enterprise admins can sign a member out of Claude everywhere with the Admin API, also in beta.
The Claude Help Center describes IP allowlisting as a public beta for Enterprise plans and Claude Platform organizations. Owners on Enterprise plans and admins on Claude Platform can set up and manage the list themselves, without going through Anthropic. The documentation notes that settings and behavior can change during the beta. None of these sources lists a separate price.
Anthropic is direct about the limits of the control: an allowlist restricts where a key works, but not for how long. The documentation recommends pairing it with a short key expiration or moving to Workload Identity Federation so you stop relying on static keys.
What the allowlist covers
Coverage differs between Claude Platform (Console and API) organizations and Claude Enterprise (claude.ai) organizations, which each keep their own list.
| Organization type | Covered | Not covered |
|---|---|---|
| Claude Platform (Claude Console) | Console UI; Messages, Message Batches, Token Counting, Models, Files and Skills endpoints; Managed Agents Agents, Sessions, Environments, Vaults and Memory stores endpoints; Admin API and Compliance API with a key that covers only this organization | Claude in Amazon Bedrock, Claude Platform on AWS, Claude Models on Google Cloud, Claude in Microsoft Foundry, and networking inside the Managed Agents runtime |
| Claude Enterprise (claude.ai) | Claude on the web, desktop and mobile; Claude Code when users sign in with a Claude account; Claude in Chrome, Office add-ins, Cowork and Remote Tools; connectors and artifacts in those apps; Admin API and Compliance API with an organization-only key | Claude Design, Claude in Slack, Claude Academy, and artifacts a user shared outside the organization |
API coverage applies whether the request uses an API key or a Workload Identity Federation token. In Claude Enterprise, a task already running in a chat can continue for up to 24 hours, but the blocked user cannot open or control it. If your proxy sends bridge.claudeusercontent.com out through a different address from claude.ai and api.anthropic.com, list that address too, because Claude in Chrome, Cowork and Remote Tools connect through it. Teams that also govern prompts with Claude inference hooks can use the allowlist as a separate network-level layer.
How to turn on the IP allowlist
The setup steps are the same for both organization types; only the settings page differs.
- Inventory every public egress address first: data centers, CI systems, offices, VPNs, proxies, and the workers of any self-hosted sandboxes, in both IPv4 and IPv6.
- Open the settings page. In the Claude Console, go to the Identity and access page. In Claude Enterprise, go to claude.ai > Organization settings > Organization and access.
- Turn on the IP allowlist toggle. Nothing changes until you save.
- Click Add entry (Console) or Add IP range (claude.ai), enter addresses or ranges separated by commas, and press Enter. Include the network you are connecting from.
- Select Allow admins to access this settings page from any IP address (Console) or Allow owners to access this settings page from any IP address (claude.ai), as Anthropic recommends.
- Click Save and confirm. Changes can take up to 15 minutes to apply.
- Watch your applications for 403
permission_errorresponses, including jobs that run only daily or weekly, and add any address you missed.
If you recently moved to API version 2026-10-21 or rely on long-running message threads, test those code paths from each production network after saving.
Entry rules and limits
- Format: a single address such as
203.0.113.5or a CIDR range such as203.0.113.0/24or2001:db8::/32. - Breadth: no broader than
/8for IPv4 or/32for IPv6. - Size: up to 1,000 entries per organization.
- Public addresses only: list the addresses your traffic leaves from, not private ranges such as
10.0.0.0/8. - Shared proxies: if you add a shared proxy address, an attacker with a leaked key or stolen sign-in could potentially use it through that proxy.
Remote staff on mobile data, home Wi-Fi, or laptops off the VPN are blocked until they connect through a listed network.
Avoiding and fixing a lockout
When you save, the settings page checks only your current address. It does not check your servers, your teammates, or the address you might use tomorrow. The escape-hatch checkbox lets an admin fix the list from anywhere, but an attacker who compromises an admin session could also change it remotely. If you set up the Compliance API, its Activity Feed records allowlist changes so you can audit them.
If you are locked out, the documentation lists four options in order: reconnect from a listed network and click Try again; ask another admin or owner on a listed network to add your address; if the escape-hatch checkbox is selected, click Open IP allowlist, then Add it or turn the list off and save; or ask Anthropic support to turn off the allowlist. In Claude Enterprise, only owners or primary owners can change the list. Access can take up to 15 minutes to return.
Related beta: sign a member out everywhere
The same release added POST /v1/organizations/users/{user_id}/sign_out for Claude Enterprise organizations, documented in user management. It signs a member out on every browser and device without removing them. It requires the write:members scope and the header anthropic-beta: admin-user-sign-out-2026-10-08. The sign-out reaches the person’s whole Claude account, and the member keeps their seat and can sign in again right away. For plan context, see how much Claude Team costs and the latest Claude Code features.
Frequently asked questions
What is the Claude IP allowlist?
It is an organization-wide list of approved public IP addresses and ranges. For each covered request, Anthropic compares the request’s public IP address with the list and denies it if the address is not allowed. It is in beta for Claude Platform (Claude Console) and Claude Enterprise organizations.
Who can turn on the IP allowlist?
In a Claude Platform organization, an organization admin or owner sets it up on the Identity and access page of the Claude Console. In a Claude Enterprise organization, owners and primary owners set it up under claude.ai Organization settings > Organization and access.
How many IP addresses can I add?
The list holds up to 1,000 entries. Each entry is a single IPv4 or IPv6 address or a CIDR range no broader than /8 for IPv4 or /32 for IPv6. Use public egress addresses, not private ranges such as 10.0.0.0/8.
What error does a blocked API request return?
Covered endpoints reject the request. The Messages API, for example, returns a 403 permission_error with the message “Your IP address isn’t in this organization’s allowed range.” People in the Console or apps see a page titled “Your current network can’t access this organization.”
What if I lock myself out?
Reconnect from a listed network, or ask another admin or owner on a listed network to add your address. If the “access this settings page from any IP address” checkbox was selected, you can fix the list from anywhere. Otherwise, ask Anthropic support to turn the allowlist off.
Sources
- Claude API Docs: IP allowlisting
- Claude Platform release notes (November 5, 2026)
- Claude Help Center: Restrict access to Claude with IP allowlisting
- Claude API Docs: User management (Sign a member out everywhere)
Image: Claude logo by Anthropic, public domain (trademarked), via Wikimedia Commons. Placed on a white background and resized. Claude is a trademark of Anthropic. FreeAnswers is not affiliated with or endorsed by Anthropic.