IPv6 Has Arrived in NetBird: What Changes with v0.71 and Who Should Roll It Out

May 18, 2026
Timo WevelsiepTimo Wevelsiep
birdhost

IPv6 Has Arrived in NetBird: What Changes with v0.71 and Who Should Roll It Out

NetBird v0.71 adds dual-stack overlay networking with per-account IPv6 prefixes: what changes, how rollout works and which caveats admins need to know.

birdhost.de Blog

Content notice: The information in this article was compiled to the best of our knowledge at the time of publication. Technical details, pricing, versions, licensing models and external content are subject to change. Please verify the information independently, especially before making business-critical or security-relevant decisions. This article does not constitute individual professional, legal or tax advice.

NetBird released v0.71.0 on May 14, 2026. The main feature is not just another dashboard toggle, but a larger architectural step: NetBird's overlay network can now run dual-stack. Every account gets its own IPv6 prefix alongside the existing IPv4 range, and peers can receive an IPv6 overlay address in addition to their IPv4 address.

For many real-world environments, that lands at the right time. Home-office networks, mobile networks and cloud backends are no longer cleanly IPv4-only. If you used NetBird from modern home networks, mobile clients or in front of IPv6-only backend resources, you often had to rely on extra IPv4 plumbing.

The important part: existing NetBird installations are not forced into IPv6 everywhere. v0.71 uses a group-based rollout model, which makes the feature easy to pilot before widening the scope.

Table of Contents

What Is New in NetBird v0.71?

The release notes list IPv6 Overlay Addressing as the headline feature. NetBird's model is straightforward: every account receives its own IPv6 prefix, default /64, configurable from /48 to /120. Peers can then receive both an IPv4 and an IPv6 address in the NetBird overlay.

The important change is not only the address. The surrounding systems follow along:

Area Change in v0.71
Addressing Per-account IPv6 prefix, default /64, configurable from /48 to /120
Rollout New accounts: All group enabled. Existing accounts: opt-in under Settings > Network
DNS A and AAAA records plus reverse PTR for overlay addresses
ACLs Existing policies apply automatically to IPv4 and IPv6
Network routes IPv6 CIDRs are supported like IPv4 subnets
Exit nodes 0.0.0.0/0 gets a matching ::/0 route for IPv6-capable peers
Domain routes A and AAAA targets stay routable through the tunnel
Client opt-out Individual hosts can stay IPv4-only with netbird up --disable-ipv6

This is cleaner than an isolated IPv6 switch. NetBird does not treat IPv6 as a special corner case; DNS, policies, routes and exit-node behavior move with it.

Why IPv6 Matters for NetBird Admins in 2026

IPv6 is no longer a future topic. Many consumer and mobile networks have been dual-stack or IPv6-prioritized for years. That matters especially for home-office and mobile clients: the laptop or phone is no longer sitting behind a simple IPv4 NAT in every case. Quite often, IPv6 is the natural path.

Backend environments are moving in the same direction. Dual-stack Kubernetes, IPv6-only cloud subnets and modern network segments are increasingly normal. A remote-access overlay that only speaks IPv4 still works, but it forces admins to add translation layers.

NetBird v0.71 closes a practical gap:

  • Mobile clients can operate more naturally in IPv6-heavy networks.
  • Backend resources with IPv6 CIDRs can be integrated without IPv4 workarounds.
  • Exit nodes can truly route "all traffic" without IPv6 leaking outside the tunnel.
  • Policies do not need to be maintained twice for two address families.

That does not mean every organization should enable IPv6 everywhere immediately. It does mean that if you run NetBird as a strategic remote-access or ZTNA component, v0.71 is more than a cosmetic release.

NetBird vs. Tailscale: Comparing the IPv6 Model

Tailscale has supported IPv6 for longer. The difference is in the model.

Tailscale uses the global ULA prefix fd7a:115c:a1e0::/48, from which devices in tailnets receive private IPv6 addresses. That is pragmatic and works well for many setups. NetBird takes a different path with v0.71: every account gets its own prefix.

Point NetBird v0.71 Tailscale
IPv6 address space Dedicated prefix per account Global ULA prefix fd7a:115c:a1e0::/48
Configurability /48 to /120 Not freely configurable
Default /64 per account Addresses from Tailscale's reserved range
Multi-tenant isolation Address space separated per account Separation primarily through tailnet/control plane
Public routability No, overlay-only No, private tailnet addresses

For a single organization without complex tenant boundaries, the practical difference can be small. For MSPs, multi-tenant deployments or organizations that intentionally separate address space per customer or environment, NetBird's per-account approach is the cleaner architecture.

This is not a "Tailscale bad, NetBird good" argument. Tailscale brought IPv6 into practical use early and robustly. NetBird is catching up with v0.71, but chooses a model that fits self-hosted and multi-tenant installations particularly well.

For a broader architecture comparison, see our NetBird vs. Tailscale comparison.

How Rollout Works

NetBird does not simply enable IPv6 everywhere in existing installations. That matters.

New accounts have IPv6 enabled for the All group by default. Existing accounts opt in under Settings > Network and select the groups that should receive IPv6 addresses.

Only peers that belong to at least one selected group receive an IPv6 address. The client also has to advertise IPv6 support. Older agents stay IPv4-only and keep working until they are upgraded.

A pragmatic rollout looks like this:

  1. Upgrade the management server and clients to v0.71+
  2. Create a small pilot group
  3. Enable IPv6 only for that group
  4. Test DNS, ACLs, exit nodes and routes
  5. Update mobile and home-office clients
  6. Expand to additional groups step by step

That lets you treat IPv6 like any other network change: start small, measure, then roll out.

What You Should Test

After the update, the question is not "Is IPv6 on?" but "Is IPv6 on where it makes operational sense?"

1. Check Client Versions

Peers should run v0.71 or newer:

netbird status

If IPv6 is enabled, netbird status also shows the IPv6 address. If you only need the IPv6 address for scripting or inventory, NetBird documents a dedicated status flag:

netbird status --ipv6

2. Check DNS

NetBird provides A and AAAA records for dual-stack peers. Reverse DNS works as well. In practice, check whether your internal tooling, logs and monitoring systems handle both address families cleanly.

dig A peer-name.netbird.selfhosted
dig AAAA peer-name.netbird.selfhosted

Adjust the zone name to your self-hosted installation. The exact name is not the point; predictable A and AAAA resolution is.

3. Validate ACLs

Existing group-to-group policies should automatically apply to IPv4 and IPv6. Still, test a few paths: SSH, RDP, HTTP or an internal API port once over IPv4 and once over IPv6.

If a policy works over IPv4 but not over IPv6, the cause is usually not a missing second policy. Look at client version, group membership, host firewall and routing first.

4. Test Exit Nodes with IPv6

If an exit node previously routed 0.0.0.0/0, NetBird creates a matching ::/0 route for IPv6-capable peers. This is where testing matters, because IPv6 leaks in "send all traffic" setups are otherwise easy to miss.

curl -4 https://ifconfig.me
curl -6 https://ifconfig.me

The IPv6 test should use the expected exit path. If curl -6 bypasses the tunnel or fails, the exit-node or client path still needs work.

5. Check Container Routing Peers

Routing peers in containers are the most important caveat in this release. NetBird tries to set net.ipv6.conf.all.forwarding=1. In unprivileged containers or locked-down Kubernetes pods, that sysctl can be read-only. The write fails silently and IPv6 forwarding remains off.

For Docker or Podman Compose, set the sysctl at container startup:

sysctls:
  - net.ipv6.conf.all.forwarding=1

In Kubernetes, the setting belongs into the pod or node configuration:

securityContext:
  sysctls:
    - name: net.ipv6.conf.all.forwarding
      value: "1"

If a routing peer has an IPv6 address but traffic does not reach the backend, this is the first place to check.

The Honest Caveats

Overlay IPv6 Is Not Public IPv6

NetBird's IPv6 addresses are overlay addresses. They are reachable inside the NetBird network, but they are not publicly routed on the internet. This matches the existing IPv4 overlay model.

If you want a peer reachable publicly over IPv6, you still need a public IPv6 address from the hoster or ISP. NetBird's overlay address does not replace that.

Mixed Fleets Stay Mixed During Rollout

Older clients remain IPv4-only. That is good for compatibility, but it also means different peers can have different connectivity profiles during rollout. Plan for that, especially with mobile clients, older Linux packages and remote sites that are rarely updated.

Individual Hosts Can Stay IPv4-Only

Not every device needs IPv6. If compliance requirements, broken kernels, embedded systems or legacy applications speak against it, use:

netbird up --disable-ipv6

That is cleaner than excluding an entire site or group from rollout.

What Else Is in v0.71?

IPv6 is the headline feature, but it is not the only relevant change.

MFA for local users: Local users without an external identity provider can enable multi-factor authentication directly in NetBird. For self-hosted setups without an external IdP, this closes an important gap.

Bring Your Own Proxy backend: NetBird added backend support for per-account reverse-proxy lifecycle. The dashboard is coming later, according to the release notes. For operators, the direction is still relevant because it moves the proxy architecture further toward multi-tenancy.

Public IPv4/IPv6 posture checks: Posture checks can evaluate public IPv4 and IPv6 properties. For zero-trust setups with network-location conditions, that is a useful building block.

Debug bundles: MTU, SSH auth config and the peer public key are now included more broadly in debug bundles. It sounds small, but it saves time in support and incident situations.

DNS upstream failover: NetBird skips failover on definitive EDE responses, reducing wrong retry paths for DNS failures.

API Changes for Integrations

If you operate your own NetBird integrations, inventory jobs or compliance checks, include the new IPv6 fields.

Object Field Meaning
Account settings ipv6_enabled_groups Groups for which IPv6 is enabled
Account settings network_range_v6 Account IPv6 CIDR
Peer ipv6 Read-only IPv6 address of the peer

If you run custom dashboards, CMDB syncs or audit exports, IPv6 should be a first-class field, not a free-text fragment in logs.

Rollout Plan for Admins

Within 7 Days: Assessment and Pilot

  • Check management server version
  • Inventory client versions
  • Define a pilot group with 3 to 5 peers
  • Enable IPv6 only for that group
  • Keep individual hosts IPv4-only with netbird up --disable-ipv6 where needed

Within 30 Days: Validation

  • Check A and AAAA resolution
  • Test reverse PTR in logs and monitoring
  • Sample ACL behavior over IPv4 and IPv6
  • Test exit nodes with curl -4 and curl -6
  • Check container routing peers for net.ipv6.conf.all.forwarding=1

Within 60 to 90 Days: Wider Rollout

  • Enable additional groups
  • Update mobile clients
  • Introduce public IPv4/IPv6 posture checks
  • Extend operating documentation for IPv6
  • Update API consumers and monitoring for IPv6 fields

What This Means for birdhost Customers

For birdhost customers, v0.71 is primarily an update topic, not an infrastructure project. We run it as managed NetBird hosting and handle management-server updates, validation and rollout planning.

That means:

  • no manual management-server maintenance
  • controlled update windows
  • validation of DNS, ACLs, exit nodes and routing peers
  • support for pilot groups and rollout strategy
  • hosting in ISO 27001 certified German data centers (certification held by the data center operator)

If you self-host NetBird, v0.71 is still a good reason to review your update process and to weigh self-operation against a managed instance at a fixed price. If your remote-access stack is business-critical, a release like this should not disappear between daily operations and "we'll do it later."

Summary

NetBird v0.71 is one of the more important NetBird releases of 2026. IPv6 Overlay Addressing brings NetBird closer to the network reality many organizations already live in: dual-stack clients, mobile users, modern home offices and IPv6-capable backend environments.

The three most important points:

  1. IPv6 is account-scoped and group-gated. Existing installations are not blindly changed.
  2. DNS, ACLs, network routes and exit nodes follow automatically. Admins do not need duplicate policies.
  3. Container routing and old clients are the main caveats. Test those paths first.

With birdhost, you get NetBird as a fully managed service: dedicated instance, updates handled by our team, hosting in Germany and personal support.

Get started now: birdhost.de


Sources

Frequently Asked Questions

What is new in NetBird v0.71?
The main change in NetBird v0.71.0, released on May 14, 2026, is dual-stack overlay addressing. Every NetBird account gets its own IPv6 prefix alongside the existing IPv4 range. DNS, ACLs, network routes and exit nodes work across both address families automatically. The release also includes backend support for Bring Your Own Proxy, MFA for local users without an external identity provider, public IPv4/IPv6 posture checks and several userspace packet filter improvements.
What does IPv6 Overlay Addressing mean in NetBird?
NetBird assigns private overlay addresses to peers. Until v0.70, that overlay was IPv4-only. With v0.71, the overlay can run dual-stack: every account gets its own IPv6 prefix, default /64 and configurable from /48 to /120. Peers can then receive both an IPv4 and an IPv6 overlay address. These addresses are not publicly routable. If you want a peer reachable on the public IPv6 internet, you still need public IPv6 on the host.
How is NetBird's IPv6 model different from Tailscale?
Both tools use private IPv6 addressing for their overlay networks, but the architecture differs. Tailscale assigns devices private IPv6 addresses from the global fd7a:115c:a1e0::/48 ULA prefix reserved for Tailscale. NetBird v0.71 assigns a separate prefix per account, default /64 and configurable from /48 to /120. For MSPs, multi-tenant environments or organizations with address-isolation requirements, NetBird's per-account model is cleaner. For simple single-tenant deployments, the practical difference is smaller.
Do I automatically get IPv6 when upgrading to v0.71?
Not across existing deployments. New NetBird accounts have IPv6 enabled for the All group by default. Existing accounts opt in under Settings > Network by selecting one or more groups that should receive IPv6 addresses. Only peers in at least one selected group and running a v0.71+ client receive an IPv6 address. Older clients remain IPv4-only until they are upgraded.
Which platforms support IPv6 in NetBird v0.71?
IPv6 support is platform-wide: Linux with kernel and userspace WireGuard, macOS, Windows, iOS, Android, FreeBSD and the netstack/userspace path for sandboxed environments. The requirement is a v0.71 or newer client. The client advertises IPv6 capability when it connects; older agents do not receive an IPv6 address and continue to operate over IPv4.
What happens to existing ACLs when IPv6 is enabled?
Existing ACLs continue to work and apply automatically to both address families. A group-to-group policy for TCP/22 applies to IPv4 and IPv6 connections between peers in those groups after IPv6 is enabled. On nftables hosts, NetBird creates a parallel IPv6 table. On iptables hosts, it uses -6-suffixed ipsets to avoid collisions with IPv4 sets. Admins do not need to duplicate policies.
What should I watch for with container-based routing peers?
IPv6 forwarding has to be enabled at the orchestrator layer, not only inside the container. NetBird tries to set net.ipv6.conf.all.forwarding=1 on startup. In unprivileged containers or locked-down Kubernetes pods, that sysctl can be read-only, the write fails silently and IPv6 forwarding remains disabled. In Docker or Podman, set the sysctl at container startup. In Kubernetes, use securityContext.sysctls or the appropriate node-level configuration.
How do I disable IPv6 for individual hosts?
Use the client flag netbird up --disable-ipv6. The client then does not request an IPv6 address, does not advertise IPv6 capability to management and does not accept inbound IPv6 traffic from remote peers. The desktop UI exposes the same option under Settings > Disable IPv6; iOS and Android have it under Advanced Settings. This is more precise than excluding a whole group from the rollout.
Should I upgrade to v0.71 immediately?
If you operate NetBird, you should test v0.71 and plan rollout in the next maintenance windows. For IPv4-only existing deployments, the update is not automatically disruptive because IPv6 must be enabled explicitly. The release also brings improvements to the userspace packet filter, DNS upstream failover, debug bundles and MFA for local users. Mobile-heavy and home-office-heavy environments benefit from the IPv6 rollout more quickly.
Do I need public IPv6 from my ISP to use IPv6 in NetBird?
No, not for overlay addressing. The IPv6 addresses NetBird assigns work inside the NetBird network regardless of whether the local ISP provides public IPv6. If you want a peer reachable from the public IPv6 internet, you still need public IPv6 on the host. For peer-to-peer connectivity, exit nodes and internal services, NetBird overlay IPv6 is sufficient.
Timo Wevelsiep

Written by

Timo Wevelsiep

Founder, merkaio

Founder of merkaio. Managed NetBird VPN hosting. Focused on network security, zero-trust architecture and scalable VPN infrastructure.

LinkedIn

Request Managed NetBird

We operate your dedicated NetBird instance including hosting, updates, monitoring and support. Tell us how many users, sites or devices you want to connect. We'll get back to you within 24 hours with a suitable proposal.

Timo Wevelsiep

Your Contact

Timo Wevelsiep

Founder, merkaio

Discuss your project with Timo

By submitting, you agree to our Privacy Policy.