01 What is this thing?
If you've ever used VOACAP, VOACAP Online, or watched a contest DXpedition plan its band schedule, you already know the engine underneath this: it's the same NTIA/ITS point-to-point HF prediction model that's been the industry standard for propagation forecasting since the 1980s. It takes a transmitter location, a receiver location, a month, and a sunspot number, and tells you — for every hour of the day and every frequency you care about — how reliable that circuit is likely to be.
voacap-skill takes that engine (via voacapl, the modern Linux/GFortran port, on macOS/Linux, or the native NTIA/ITS Windows engine on Windows) and wires it into Claude Code as a "skill" — a capability Claude can reach for on its own. Instead of filling out a web form, you just ask, conversationally, in the same terminal where you're already working:
What's the best band to reach Tokyo from Columbus, Ohio around 0300Z in October?
Claude runs an actual VOACAP prediction in the background and comes back with a real answer — not a guess, not a canned response, but output from the same math that predicts circuit reliability for professional HF networks.
REL — Reliability
The predicted fraction of days in the month your circuit hits the SNR you need. 0.7–1.0 is a solid, workable circuit; below 0.3 and you're fishing.
SNR — Signal-to-noise
How far your signal is expected to sit above the noise floor, in dB, for the mode/threshold you asked for.
MUF — Maximum usable freq.
The highest frequency the ionosphere is predicted to support on that path at that hour — the ceiling you're working under.
02 Installing it
You'll need Claude Code. Inside a session, run:
/plugin marketplace add Reid-n0rc/voacap-skill
/plugin install voacapThat's it — VOACAP builds itself the first time you use it. A manual install (no plugin manager) is also available — see the README.
03 Using it
The natural way to use this is to just talk to Claude — describe your station, your target, and roughly when, and let it figure out the right VOACAP parameters. It'll ask for anything important it's missing, like an up-to-date sunspot number.
Under the hood it's a thin wrapper around one script that prints one row per UTC hour, with the best-performing frequency for that hour and the REL/SNR for every frequency you tried. Full flag reference is below in Section 5, pulled straight from the skill's own docs — it can't drift out of date.
04 Fun things to try
Is 20 meters going to be open to Japan for CQ WW SSB this year? I run 100 watts to a dipole from Seattle.
Contest planning: check reliability across the contest window and get an honest read on whether your run band will actually support the path before you build your whole weekend around it.
What's the best band for a grayline QSO between my station in Denver and Cape Town, South Africa?
Grayline propagation is a favorite low-band trick — ask Claude to scan the sunrise/sunset hours and compare 40m, 30m and 20m reliability along the path.
We're planning a DXpedition to a small island. Reachable from EU and NA — which bands should we run at each UTC hour to work both continents?
Run the same TX location against two RX targets side-by-side and build an actual band-by-hour operating schedule, the same way real DXpeditions plan their logs.
For the Tennessee QSO Party this weekend, what's my best band from Chattanooga to reach stations in the Midwest and West Coast at 2100Z?
State QSO parties live and die by band changes over a single afternoon — ask hour-by-hour and know when to move before the pileup tells you.
How much does bumping my power from 100W to 1kW actually change my reliability on 40m at night to the East Coast?
A great gut-check before spending money on an amplifier — run the same circuit twice with --tx-power-kw changed and compare REL side by side.
Given SSN=150 vs SSN=20, how differently does 10 meters behave on a transatlantic path?
A quick, concrete feel for where you are in the solar cycle — run the identical path at both sunspot extremes and watch the high bands come alive (or die).
05 Full command reference
This section is fetched live from the skill's own
SKILL.md
on every page load — so it always matches whatever is actually on
main, even if this page itself hasn't been redeployed.
Loading current docs from GitHub…