Who runs NeedPorts

Created by Vast hosts, for Vast hosts.

NeedPorts exists because public inbound connectivity is one of the most frustrating parts of running GPU hosts, self-hosted services, and remote workloads. A service can be healthy locally, the machine can have outbound internet, and still nobody can reach it from the public internet.

The short version

NeedPorts is built from the operator side of the problem. We understand the pain of hosts that should be useful but are hidden behind CGNAT, provider filtering, missing public IPv4, or network setups where normal router-style port forwarding is not available.

That is why the product is intentionally simple: a stable public endpoint, a dedicated forwarded port range, and an outbound tunnel from your machine back to NeedPorts.

Why that matters

Many networking products are built for general enterprise VPN use cases. NeedPorts is narrower. It is designed for people who run machines and services that need to be reachable from outside:

Vast/GPU hosts

GPU machines, inference endpoints, dashboards, SSH, and workload callbacks that need stable inbound ports.

Self-hosted services

APIs, admin panels, game servers, web services, and homelab tools that work locally but time out externally.

CGNAT and filtered networks

Networks where you can connect outbound, but cannot receive public inbound connections directly.

Built for practical operations

NeedPorts focuses on boring, useful infrastructure details: one-command setup, systemd service management, auto-reconnect, dedicated assigned ports, and clear install/status commands. The goal is not to make customers learn tunneling internals. The goal is to make the host reachable and keep it reachable.

NeedPorts is not a privacy VPN and it is not a magic fix for an application that is not running. It helps when the real blocker is public inbound reachability and the machine can make outbound connections.

What we optimize for

Simple setup

Paste one install command, receive your assigned public endpoint and port range, then verify the client is connected.

Clear ownership

Each customer gets an assigned range and token. The model is easier to reason about than shared, rotating, single-port workarounds.

Useful support

When something fails, we care about the actual path: local service health, outbound tunnel state, public endpoint state, and assigned port reachability.

Responsible use

NeedPorts should be used for legitimate services you control, subject to provider policies and applicable laws. We built it to solve real reachability problems for hosts and services, not to hide abuse or bypass rules.

Want to see the technical model?

Read How NeedPorts Works for the plain-English explanation of public endpoints, outbound tunnels, and dedicated forwarded ports.

Start with a Vast/GPU host Start with self-hosting