Desktop traceroute map · Windows, macOS, Linux · v0.5.0-beta.5
See where your packets actually went.
Route Map is a free, open source visual traceroute for Windows, macOS and Linux. It runs the trace on your own computer and draws every hop on a world map: each router is placed by its operator's own hostname before any IP database is asked, every placement the round-trip time makes impossible is rejected, and each hop says where its location came from.
Also: Windows Installer Windows Zip macOS Apple silicon macOS Intel Linux Debian, Ubuntu (.deb) Linux Fedora, RHEL, openSUSE (.rpm) Linux AppImage Linux tar.gz. v0.5.0-beta.5, checksums and install steps.
Free and open source, AGPL-3.0 github.com/osintph/routemap

A saved trace from Manila to heise.de, drawn by the app itself. The first hops, the home network and the ISP, are documentation addresses.
How it places a hop
Most of what a visual traceroute shows depends on where it puts each router, and that is where most of them go wrong.
Hostname first
Operators name routers after their sites: ffm-bb2-link.ip.twelve99.net is
Arelion in Frankfurt, and sto, hnk, sng and
prs are Stockholm, Hong Kong, Singapore and Paris. For backbone routers an IP
database often knows only the country, or the place where the address block was registered.
Route Map reads the names first, with CAIDA's Hoiho rules and a table of carrier site codes,
and falls back to an IP database only when the name says nothing: DB-IP Lite on your
machine, or RIPEstat online.

Checked against physics
Light in fibre covers about 200 km per millisecond, so each millisecond of round trip allows about 100 km: a hop that answers in 5 ms is within about 500 km, plus 300 km of slack for an approximate origin. Any placement farther than that is rejected, whatever its source, and the reason is shown. So is a database answer that would need a detour the round trip did not pay for. A destination that answers faster than its registered place allows, as anycast networks do, is placed where the round trip past the last placed hop reaches, and labelled "served from near" that city.

Hop 8 answered from two routers. The name of one says Marseille, 10,849 km from Manila; its fastest probe came back in 58 ms, which allows at most 6,130 km, so that place is ruled out and the other router's Singapore is used.
Nothing guessed, nothing dropped
Every hop carries its source, and a hop the database knows only to a country is drawn hollow. A hop that cannot be placed stays in the table and in the Unplaced panel, with the reason it is not on the map.

Beyond where the hops are
Beside the map: the networks your packets crossed (the AS path, from DB-IP Lite ASN on your machine) with an RPKI badge for each, the countries they went through, what RIPE's route collectors see for the destination prefix, its BGP activity over the last 48 hours, and a typical latency for the same trip from RIPE Atlas anchors. Lines follow the great circle and warm with the round-trip time each step added; switch to the globe to see the path as the packets flew it. Trace again and the app compares the two runs.


No submarine cables or exchange points yet: the cable data is sold under licence, and PeeringDB's terms do not allow bundling its list. The FAQ has the details.
On your machine, with your tools
The trace runs on your computer with its own tracert, traceroute or
mtr, and the status bar says which, with which flags, and which lookups answered.
Hops appear on the map and in the table as the trace runs; click a marker to find its rows, or
a row to find its marker. Paste a trace someone else ran, say where it came from, and it is
mapped the same way. Export a PNG map, the full trace as JSON, or a PDF report with every hop,
its source and the raw tool output, from the window or the command line.
The same engine powers the Route Map tab in OSINTPH's FalconEye.
Watch keeps tracing, mtr style: every hop once a second, with loss, latency and jitter per hop, a ping plot for the hop you select and the destination, and a note whenever the path changes. Loss that only a router's own replies show is marked as rate limiting, a computer that sleeps leaves a gap, not loss, and a change of its clock is not a gap. It never probes faster than mtr's defaults and stops by itself after an hour; export the session as PDF or JSON and compare it like any trace. How it works.
Paths finds the paths a load balancer spreads your traffic over; some may stay hidden. It uses Paris traceroute and the multipath detection algorithm's stopping rule, and gives each path its own latency and loss: on the map one colour per path, in the table a row per router where they part ways. It stops after 5 silent hops, as scamper and RIPE Atlas do, then asks the target once more at full TTL on each path found. On Windows it probes over TCP; a hop that answers no TCP probe gets one ICMP echo, and the heading says whether the routers there do not answer TCP or the target answers ICMP while its TCP service stays silent. At most 20 probes a second and 1,500 in all. Reverse trace asks a RIPE Atlas probe near the target to trace back to you, with your own Atlas key, and shows both directions side by side, aligned by network. It publishes your public IP address as the measurement's target, so it asks first. Traces, Paths and Watch run over IPv4 or IPv6. Paths, reverse trace.
DNS timing. Every trace times the lookup of its target, A and AAAA, through your system resolver and any resolvers you add, and shows it with the trace: the address traced, each answer's time, whether the resolver says it is authoritative, and its TTL. Route Map resolves the name once and traces that address, so the address shown is the one traced. Plain DNS is not encrypted: only the resolvers you list are asked, and with Online lookups off only your system resolver. DNS timing.
Ports tries one TCP connection to each port you choose on the traced address, 443, 80, 22, 25, 53 and 3389 by default and at most 32, and reports each as open with its connect time, closed or filtered, with the TLS handshake time for 443. It is a reachability check for one target, not a port scanner: no ranges, four at a time. Next to Paths, an open port 443 says the target is up and the routers in between are what is silent. Port tests.

A remote agent is a server you already reach with ssh. Route Map runs one command there for each trace over your own ssh client, with no daemon on the server and no keys kept by the app, and maps the result from the server: Trace, Paths or Ports from a data centre or another country. The server needs routemap-engine for the user you log in as. Its first hops are the provider's own routers, so they are hidden by default. Remote agents.
Behind a VPN or SASE client, Route Map says when the route goes through a tunnel adapter, warns when the origin is probably the tunnel's exit, checks TCP Paths for a local proxy, and lets you choose the adapter to probe from. See the guide.

The source labels
hoiho- A CAIDA Hoiho rule for that operator matched the hostname and names the city.
site-code- The hostname carries a carrier's own site code, from the bundled table.
ip-db- An IP database placed it, DB-IP Lite on your machine or RIPEstat online (the label says which): a hint, since backbone addresses often sit where the prefix was registered.
ip-db, country only- Only the country is known; hollow marker, kept out of the map's fit.
anycast- The destination answered faster than its registered place allows, so it is served from near you: placed where the round trip past the last placed hop reaches. An inference.
inferred from neighbours- No source placed it, but the hops either side are in one city; hollow marker. An inference.
local- Your network or your ISP's carrier NAT; never looked up anywhere.
What leaves your machine
| What | To | When |
|---|---|---|
| Public router hostnames | CAIDA Hoiho | each trace, unless switched off; cached 30 days |
| Public hop addresses DB-IP Lite does not place | RIPEstat | each trace, unless switched off; cached 30 days |
| Public addresses, prefixes and AS numbers, for route details | RIPEstat | after each trace, unless Online lookups are off |
| Two country codes and an anchor name | RIPE Atlas | after each trace, for the typical latency; public data |
| A download request | DB-IP | only when you choose to download DB-IP Lite |
| Reverse DNS queries | your own resolver | for public hops without a name |
| ICMP probes; on Windows TCP connection attempts to port 443, plus one ICMP echo per hop that did not answer TCP | the target and the routers on the way | each Paths run, at most 20 a second and 1,500 in all; the same lookups as one trace |
| The target's name | your system resolver, and the resolvers you add in Settings › Probes (none by default), nowhere else | each trace; plain DNS is not encrypted; with Online lookups off, only the system resolver, which the trace needs |
| One TCP connection attempt per chosen port (at most 32), and for TLS ports a TLS ClientHello with the target's name | the traced address | only when you click Ports, or turn on Test these ports with every trace |
| One ssh connection, with your own ssh client | the server you named in Settings › Agents, which sends the probes | only when you trace from an agent; "From the agent's public IP" makes the server look itself up at RIPEstat, so RIPEstat sees the server's address; "City" makes no lookup |
| Your public IP | RIPEstat | only while no origin is set in Settings |
| Your adapter list, local addresses and gateways | nobody: read locally, never sent | in an export only if you tick the box in the Export dialog |
| Target, key, your network's AS | RIPE Atlas | only for an Atlas trace, with your own key; public |
No account, no telemetry, no automatic update check. One switch, Online lookups, turns every lookup off. Private, CGNAT and reserved addresses are never looked up. The full list: privacy.
Support the project
Route Map is free. Donations pay for code signing, hosting, the RIPE Atlas probe, and maintenance time.