Tixati proxy settings: trackers, peers or both, and the DHT traffic that never uses the proxy

Tixati's help page for Settings > Network > Proxy is short and unusually candid: "A SOCKS4/4a/5/HTTP proxy server can be used for tracker, peer, or both types of connections. It is not possible to use a proxy for DHT UDP packets, and using a proxy for peers may prevent incoming connections from being received." Everything on this page follows from that sentence. Proxy peers and trackers, then turn DHT off, and prove both with a count.

We did not run Tixati in our lab. The quotes are Tixati's; the measurement is a procedure for your machine.

#What the sentence means on our proxies

• DHT goes direct, always. With DHT on, Tixati sends DHT packets from your own address whatever the proxy settings say. That is the leak to close.
• The rest of UDP is unconfirmed. Whether Tixati relays UDP trackers or its own UDP peer connections through a proxy, its documentation does not say. It matters little on our gateways: treat our SOCKS5 as TCP only, since we do not advertise UDP relay. An HTTP proxy carries no UDP at all.
• Fewer peers. Behind the proxy you connect out to peers that accept incoming connections; they cannot connect in to you.
• No encryption. The proxy relays traffic as it is. Anything sent outside it shows your own IP.
• Billed both ways on a meter. Residential and over-allowance shared traffic is billed per GB, and seeding pushes the payload through again. How traffic is counted is in the terms.

Tixati's site advertises no ads and no spyware, which is welcome, but a proxy setting is not a privacy setting. The qBittorrent project's wiki says it plainly to its own users: for copyright concerns, use a VPN. Our proxies are for lawful swarms, and unauthorised sharing through them is not allowed by the allowed-use policy.

#Set it

1. Open Settings > Network > Proxy.
2. Choose SOCKS5 and enter the host and SOCKS5 port from your order, or HTTP with the HTTP port.
3. Apply the proxy to both tracker and peer connections. Tracker-only proxying hides your address from the tracker operator while every peer sees it on the data connections.
4. Credentials: if your version offers username and password fields, use them. If it does not, add your public IP to the order's IP allowlist and leave the client without credentials.
5. Turn DHT off. Tixati 3.43, released June 19, 2026, brought major DHT changes, so look for the switch in the version you run rather than in an old screenshot.

#Measure: two runs, DHT on and DHT off

Start a lawful torrent with an HTTP tracker, such as a current Debian or Fedora image, and do two short runs, restarting Tixati between them.

Linux:

PID=$(pgrep -i tixati | head -1)
PROXY=203.0.113.10:1080
ss -Htnp state established | grep "pid=$PID," \
  | awk -v p="$PROXY" '$4 == p { via++ } $4 != p { direct++ } END { print "via proxy:", via+0, "direct:", direct+0 }'
ss -Huanp | grep "pid=$PID,"
sudo timeout 60 tcpdump -nni any "udp port LISTEN_PORT" 2>/dev/null | wc -l

Windows (PowerShell, with the PID from Task Manager):

Get-NetTCPConnection -OwningProcess 1234 -State Established | Group-Object RemoteAddress | Select-Object Count, Name
Get-NetUDPEndpoint -OwningProcess 1234

and in Wireshark, udp.port == LISTEN_PORT over the same minute.

Write both numbers down. The difference between them is DHT, measured, and the second run is the configuration to keep. Check the exit IP with the curl line from our proxy checker so you know which address the swarm should see.

#Reconcile the client with the meter

Tixati's strength is its peer and bandwidth views, which show how much each peer sent and received. Use them. On a residential order, note the traffic balance in the dashboard before a seeding session and again after it, and compare the difference with the total Tixati reports. Proxied payload plus protocol overhead should account for the difference, and a figure far below Tixati's total means some peers were not proxied.

At checkout prices, a 3 GB release downloaded and seeded to a ratio of 1.0 is 6 GB through proxied peers: $27.60 on residential. Seeding it to 3.0 is 12 GB, $47.40. A datacenter IP is $1.90/IP (dedicated) a month with no bandwidth cap, bounded by port speed and fair use. Peer-to-peer traffic for lawful content is fine on residential and on dedicated ISP and datacenter IPs, which carry no traffic meter. The cost calculator compares them at your volume.

The torrent client comparison places Tixati with aria2 and Transmission as the clients whose UDP goes direct, and BiglyBT as the one with the most complete leak checklist.