Should I run a Tor middle relay, bridge or exit?

What is a Tor guard or middle relay?

A guard relay is the first hop in a Tor circuit; a middle relay is the second. The Tor Project calls both non-exit relays. A guard has to be stable and fast, at least 2 MByte/s each way, or it stays a middle relay.

Non-exit relays rarely draw abuse complaints, because they only pass encrypted traffic between relays and never connect out to a website. They need little upkeep. The main catch: every relay's IP is in the public directory, so some services that misread Tor may block it. A non-exit relay is the lowest-risk way to donate real bandwidth, and a Ravenquay berth has far more than one needs.

What is a Tor bridge?

A bridge is a relay that is not listed in the public Tor directory, so a government or ISP that blocks every public relay IP cannot block it as easily. Bridges help people connect from censored networks.

The Tor Project says bridges are the most needed relay type right now. A plain bridge adds the obfs4 pluggable transport, which disguises the connection; WebTunnel wraps it to look like ordinary HTTPS traffic to a website you host. Both run on modest hardware, from 1 Mbit/s up. Because a bridge IP is not published, it is far less likely to be blocklisted by other services. For the wider picture, read what censorship-resistant hosting is.

What is a Tor exit relay, and why is it different?

An exit relay is the final hop: it connects the circuit to the open internet, so the website sees the exit's IP, not the user's. That is why exits carry the greatest legal exposure of any relay. Abuse complaints about exit traffic come straight to the operator, through the host named in the WHOIS record.

The Tor Project is blunt: do not run an exit relay from your home. It advises running exits on a server dedicated to the purpose, in a datacenter, with its own IP address, and never mixing your own traffic with exit traffic. A rented whole machine is built for exactly that. See how far an exit scales on a fast port in how much bandwidth a Tor exit relay uses.

Which relay type does the Tor network need most?

The Tor Project's support portal ranks the need: bridges first, because they reach users in censored regions; then exit relays in diverse locations, which carry the highest legal risk; then fast guard relays, which are valuable with minimal legal risk.

Network diversity matters as much as raw capacity. The Tor Project tells operators to spread relays across autonomous systems and countries, and to avoid a handful of large networks that already carry many relays. Ravenquay's four ports of call, Reykjavik (Iceland), Zurich (Switzerland), Bucharest (Romania) and Amsterdam (Netherlands): every berth in every port of call, at one price, add location diversity on their own.

Which relay type runs best on a Ravenquay berth?

All three are welcome: Tor exit relays are welcome, along with middle relays and bridges. The choice is yours, but a rented offshore machine suits an exit especially well, because the Tor Project asks for a dedicated box with its own IP, kept apart from your own traffic.

• A bridge or middle relay runs comfortably on the smallest berth, Harbor 8 with its 1 Gbit/s port.
• An exit runs on any berth; one dedicated IPv4 means up to eight tor instances.
• A high-capacity relay fits Harbor 10G and its 10 Gbit/s port.

Ravenquay keeps no logs and shuts outgoing port 25, which matches the Tor Project's default exit policy. Full prices are on the Ravenquay pricing page; setup details are in the Tor relay requirements.

How many hosts even allow exits?

Not many say so plainly. In Ravenquay's survey of 34 crypto-friendly dedicated-server hosts on 30 October 2026, only 4 allowed Tor exit relays, 1 more allowed them with notice, 8 banned them, and 21 stated no policy. A middle relay or bridge is usually fine even where an exit is not, but a host that welcomes all three removes the guesswork.