About & credits
What this is
Aerodrills is a voice-driven ground trainer. You rehearse a flight in a chair: it reads you a short setup, asks you one question, and you answer out loud. It then marks the answer against what a good one needed — what you covered, and what you left out. Each scenario is written three times over, once per level: Beginner tells you what has gone wrong and asks for the first action, Standard gives you the symptoms and the whole response, and Checkride adds a constraint that changes the answer and wants the specifics.
Safety
Aerodrills is a ground training aid. It is not a flight planning tool, not a substitute for instruction, and must never be used in an aircraft or while driving. All aircraft data are training approximations — verify against your POH.
Every scenario is written from published sources — the PHAK, the Airplane Flying Handbook, the AIM, the ACS, and the aircraft’s POH — and each one carries the citations it was written from. What it has not had is a CFI sign-off. No instructor has reviewed and approved this library, and 20 of the 150 scenarios carry a citation saying outright that an ACS task code on them still needs verifying. That is 450 scripts unreviewed, counting all three levels. Check anything here against the source before you take it into a lesson.
It is not instruction, and nothing you fly here logs toward a certificate, a currency requirement, or any other credit.
Where your runs are kept
Demo mode — no account, and no database on this deployment: a finished run is kept in this browser only, and clearing site data removes it. The flying and the scoring are real.
In practice: a run will not follow you to another device or another browser. The same warning is repeated on the debrief and on your progress page, because those are the two screens where somebody would otherwise assume there is an account behind them.
Microphone and speech
Aerodrills does not record or store your audio. That is true in every configuration; only the transcript is kept. Recognition uses your browser’s built-in speech engine. In Chrome and Edge that engine sends the few seconds you held the key to Google for transcription; Safari sends it to Apple, and on recent versions can do it on-device. Aerodrills itself never records or stores your audio — only the text ends up in your transcript.
On iOS, wired headphones avoid a microphone-routing quirk that can otherwise leave the recogniser listening to the wrong input.
You can decline the microphone entirely and type your transmissions instead. The scenario, the timing, and the scoring are identical.
Voice credits
On a deployment with a speech key, each line is spoken as you reach it by a text-to-speech vendor — OpenAI, or ElevenLabs — in the voice you pick from the panel. With no key the lines are read by your own device’s voices, and where it has none, or where you pin them in the voice panel, by the clips committed alongside the app in public/audio and rendered offline with Piper.
Those clips are what this credit is for. The instructor is Piper’s Kathleen voice, released under CC0. ATC and the passenger are LibriTTS speakers, licensed under CC BY 4.0.