Feedback: v0.74.0

[mlsmaycon]

Before posting

• I searched existing discussions and issues for similar requests.
• I checked the documentation to confirm this is not already supported.
• This is a product idea or enhancement request, not a support question.
• I removed or anonymized sensitive details from examples and screenshots.

Product area

Dashboard / Admin UI, Management service / API

Problem or use case

Agentic workloads demand tighter control, so organizations are turning to virtual keys to gate access to company-controlled endpoints and cap costs. But this introduces its own problems. Managing the keys is a burden, and it doesn't solve the classic secret-sharing problem: people can leave the company with these keys—and plenty of others—still in hand.

Proposed solution

New Feature: Agent Network
This release introduces Agent Network, a per-account LLM gateway that gives people and agents keyless, identity-based access to LLM APIs and internal resources over the tunnel. It is built on top of the existing reverse proxy and private services, so the transport is still NetBird's WireGuard overlay and the identity model is still your IdP. Have a tunnel, get access; no tunnel, no access.

Agents point at a tunnel-only endpoint instead of the provider's URL. NetBird holds the upstream provider key server-side, injects it per request, and ties every call to a real identity from your IdP. Client-supplied auth headers are stripped before the request is forwarded, so a hardcoded key never reaches the provider. The whole thing is default deny: nothing reaches a provider until a policy explicitly allows it.

Alternatives or workarounds considered

No response

Community impact and priority

• Every org with AI access

Examples from other tools or products

No response

Security, privacy, and compatibility considerations

No response

Implementation ideas

No response

Are you willing to help?

Yes, I can submit a PR if the approach is accepted.

Additional context

No response

[Discolor3182]

Hello,

I upgraded to 7.4.0 and getting this error when running docker compose up -d

Error response from daemon: failed to set up container networking: driver failed programming external connectivity on endpoint netbird-proxy (a1665292b6ca1eb7ee62177a2a8230f142471ee71b141deef6b7986b1b6f860f): failed to bind host port 0.0.0.0:51820/udp: address already in use

Any ideas what is causing this?

Thank you.

[mlsmaycon]

is possible that you have another process using the :51820 port?

[mertcangokgoz]

My agent network settings are open and the necessary updates have been made. The options in Log Collection are not working. There is a small "object object" situation in the notification.

[mlsmaycon]

Could you pls share logs from your netbird-server when the issue happens?

[mertcangokgoz]

Could you pls share logs from your netbird-server when the issue happens?

There are only the log lines below

2026-07-01T23:39:49.177Z ERRO [accountID: d81qjun91ggs73bporog, userAgent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/149.0.0.0 Safari/537.36 Edg/149.0.0.0, userID: CiRkYTY0OWFjNS02M2ViLTQxZWItOWUxNi00ZWRhYjRlNjY0YzUSBWxvY2Fs, context: HTTP, requestID: d92qahf91ggs73du9bs0] shared/management/http/util/util.go:85: got a handler error: get agent network settings: agent network settings for account d81qjun91ggs73bporog not found
2026-07-01T23:39:49.177Z ERRO [context: HTTP, requestID: d92qahf91ggs73du9bs0, accountID: d81qjun91ggs73bporog, userAgent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/149.0.0.0 Safari/537.36 Edg/149.0.0.0, userID: CiRkYTY0OWFjNS02M2ViLTQxZWItOWUxNi00ZWRhYjRlNjY0YzUSBWxvY2Fs] management/server/telemetry/http_api_metrics.go:210: HTTP response d92qahf91ggs73du9bs0: PUT /api/agent-network/settings status 404

[mlsmaycon]

it seems like your netbird-server is not up to date.

You can run:

docker compose pull
docker compose up -d --force-recreate netbird-server

[mertcangokgoz]

it seems like your netbird-server is not up to date.

You can run:

docker compose pull docker compose up -d --force-recreate netbird-server

If my server version is not up to date, how am I able to add a provider without any problems? This situation actually occurs when a LiteLLM addition is made, the proxy host it provides was not added to the reverse proxy(automated), and when I deleted it afterwards this situation occurred.

So actually, settings were tried to be opened for a provider that doesn't exist. In short, if no provider is added, it might be better for the log settings section to be disable.

[mlsmaycon]

I see that the steps you shared are clear. Thanks, we will have a look and fix it for the next release.