Why self-host NetBird on a VPS?
NetBird connects your devices through encrypted WireGuard tunnels and gives you a central place to manage access. A self-hosted deployment can bring a laptop, home server and cloud application into a private network that you operate.
Version 0.83.2 was released on 20 November 2026. It includes a client hostname fix and updates the self-hosted installer's Traefik image to 3.7.18. This walkthrough uses that installer release with specific server and dashboard images. Read the release notes.
Reach private services
Connect an administrator's laptop to an application's private address. Keep the application's own authentication and host firewall rules in place.
Control device access
Organise peers into groups and allow the connections each group needs. Review permissions when people, devices or workloads change.
Operate your own server
Maintain the coordination service, account access and backups. Plan how you will recover management access during an outage.
Management distributes peer information and policies. Signal helps peers negotiate connections, while a relay can carry encrypted traffic when direct connectivity fails. The dashboard and these services have different roles from the applications running on your peers. Review the architecture.
Choose a VPS and prepare DNS
The official minimum is a Linux VM with one CPU and 2 GB RAM. Our starting suggestion is a small Cloud VPS with room for system updates, logs and backup work. Add resources as your peer count and measured relay traffic grow. This is sizing guidance rather than a performance benchmark. Review the server requirements.
Use a fresh Linux VM and an administrative SSH session. Use an account that can run Docker commands directly. This guide assumes Docker Engine, the Compose plugin, curl and jq are installed. Follow our Docker installation guide first if needed.
Create an A record for a hostname you control, such as netbird.example.com, pointing directly at the VM's public IPv4 address. Publish an AAAA record only when the IPv6 address reaches the same service. Replace the example hostname throughout this guide with your own before running checks.
docker version
docker compose version
command -v curl jq
getent ahosts netbird.example.com
sudo ss -ltnup
HYEHOST Cloud VPS is available in Wolverhampton, Ashburn and Auckland. Choose the region near the users who may need the relay, and compare its transfer allowance. Leave an independent SSH or console recovery route available while changing network access.
Check the public ports before installation
| Port | Purpose | Action |
|---|---|---|
| TCP 443 | HTTPS and certificate validation | Allow inbound to the VM |
| UDP 3478 | STUN connectivity discovery | Allow inbound to the VM |
| TCP 80 | Traefik's HTTP listener | Keep free on the host |
Built-in Traefik validates certificates through TCP 443 using TLS-ALPN-01. Public TCP 80 is optional for this process. Check host and upstream firewalls, and resolve listener conflicts before starting. Check the official server requirements.
Client devices establish outbound connections and use NAT traversal. Follow the client firewall guidance for restrictive networks, substituting your own server endpoints for the hosted service's endpoints. Review client firewall behaviour.
Run the release-specific installer
Create a dedicated directory. Download the installer to a file and inspect it before execution. The following command selects the 0.83.2 server and v2.98.0 dashboard images.
mkdir -p "$HOME/netbird"
cd "$HOME/netbird"
umask 077
curl -fsSL --output getting-started.sh \
https://github.com/netbirdio/netbird/releases/download/v0.83.2/getting-started.sh
bash -n getting-started.sh
less getting-started.sh
NETBIRD_SERVER_IMAGE=netbirdio/netbird-server:0.83.2 \
DASHBOARD_IMAGE=netbirdio/dashboard:v2.98.0 \
bash getting-started.sh
Enter your real domain and certificate contact email when prompted. Choose option 0 for built-in Traefik. Keep the optional NetBird public reverse proxy disabled for this private-peer setup. CrowdSec options can remain disabled if offered. Follow the installer prompts.
The generated docker-compose.yml, config.yaml and dashboard.env describe your deployment. Read the image tags and exposed ports before adding clients. Keep these files and their secrets private. The combined server configuration replaces older instructions built around separate management and relay files. Review generated configuration.
docker compose config --quiet
docker compose config --images
docker compose ps
docker compose logs --tail=100
Create your first administrator promptly
Open your HTTPS hostname and check the certificate and address bar before entering credentials. Complete the initial setup immediately. The setup page is available while the installation has no users, and the first account becomes the owner.
Use a unique password and retain recovery information securely. Add ordinary users with the permissions they need. The embedded identity provider supports local accounts, with Owner, Admin and User roles. External identity providers can also be configured when required. Review local users and roles.
For local accounts, enable Local MFA in Settings, Authentication and save the setting. Check TOTP enrolment and retain recovery access. Configure local MFA.
Connect a laptop and one target server
Install NetBird on a laptop and one target server using the instructions for each operating system. Check that the client daemon is running. Open the client installation instructions.
On Linux, an administrator can connect the client to your self-hosted management URL.
sudo netbird up --management-url https://netbird.example.com:443
sudo netbird status
sudo netbird status -d
Follow the authentication URL and sign in to your own installation. Verify the management address and peer identity in the output. The official CLI documentation covers connection and status options. Check the client commands.
Current clients enforce profile ownership. Using sudo gives root administrative access, while a normal local user has access according to the profile they own. Keep that choice consistent when checking a headless server. Review profile permissions.
Replace broad access with useful policies
A new account starts with a Default policy that permits broad peer connectivity. Create groups for your intended access, such as administrator devices and application servers. Allow the required direction, protocol and destination port, then disable the broad default once your replacement rules are ready.
For example, an administrator group may need TCP 22 to an application-server group. Test that connection and test a device outside the allowed source group as well. Keep recovery access available while changing policies. Review groups and access policies.
Then check the application itself. A successful tunnel does not configure an SSH account, database password or listening address. Keep the service bound appropriately and review any public firewall rule that still exposes it.
Check whether peers connect directly or through a relay
Run the detailed status command on a client and inspect the intended peer's connection type and handshake. Relayed traffic remains encrypted, but its path adds an intermediary and uses relay capacity. A relay can be the expected outcome of restrictive NAT or firewall behaviour. Diagnose relayed connections.
Budget for the traffic the relay actually carries. Large file transfers or frequent backups can make its region and transfer allowance significant. Measure representative use before moving several busy workloads onto a small VPS. HYEHOST's private LAN can support suitable local VM traffic, while cross-region paths need their own transfer and latency planning.
Back up configuration, secrets and persistent state
Retain the generated deployment files and the server data at /var/lib/netbird in the persistent volume. Include any secret store supplying the authentication secret or database encryption key. For the built-in Traefik setup, retaining certificate state also helps recovery. Stop the relevant services for a consistent copy of local databases, or use an appropriate database backup method.
Protect backups as credentials and test an isolated restore with the intended software versions. Record the required hostname, storage locations and recovery steps. Follow the official procedure for the components you enabled. Review backup and restore requirements.
Update deliberately and verify connections
Record the running image versions, take a verified backup and read the target release notes. Change explicit image tags to your chosen compatible releases before pulling. Pulling images keeps the configured tags, so it will not advance a fixed tag automatically.
From the installation directory, update the server and dashboard after reviewing their versions.
docker compose config --quiet
docker compose pull netbird-server dashboard
docker compose up -d --force-recreate netbird-server dashboard
docker compose ps
Check owner login, client registration, allowed and denied traffic, and relay connectivity afterward. Traefik has its own update process. Recovery may require compatible pre-upgrade state as well as previous images. Follow the upgrade procedure.
Common NetBird setup problems
The dashboard will not open
Check DNS, both address families, listener conflicts and inbound TCP 443. Review the Traefik and server logs for certificate or routing errors.
A joined peer cannot reach an app
Check its source and destination groups, policy port, application listener and host firewall. Test the exact service address and port you intended to allow.
A connection stays relayed
Inspect detailed status, outbound UDP restrictions and NAT behaviour. Measure the resulting path before changing network-wide rules.
Capture a timestamp and the affected peer when investigating. Redact private addresses and credentials before sharing logs. Keep the deployment files intact while comparing configuration with the documentation.
Self-hosted NetBird VPS FAQ
Can I self-host NetBird without an external identity provider?
Yes. The current setup supports local accounts. Create the first administrator through the initial setup page.
Which ports does this NetBird server need?
Allow TCP 443 and UDP 3478. Keep host TCP 80 free for built-in Traefik.
Why do restrictive policies still allow every peer?
The default All-to-All policy may still be enabled. Test replacement rules, then remove that broad policy.
Does all NetBird peer traffic pass through my VPS?
Peers try direct connections. Relayed sessions use the relay path, so monitor its transfer and load.
What should I back up on a self-hosted NetBird server?
Retain deployment configuration, secrets and the management data directory. Test a consistent backup on the intended restore version.
Which HYEHOST region should I choose for NetBird?
Choose Wolverhampton, Ashburn or Auckland to suit your users and expected relay traffic. Compare the selected region's transfer allowance and recovery requirements.

