AetherPilot reads your email, decides what each message means for your work, turns the actionable ones into a checklist you approve, and measures the administrative hours that never showed up anywhere.
Every professional who bills by the hour has two kinds of time. Billable time is easy to see — it is on the timesheet. Then there is the other kind: reading email, working out what matters, filing notices, setting up new clients, keeping the file straight.
That work is real, it takes hours, and it appears nowhere. It is never measured, so it is never managed, and it quietly eats the day. AetherPilot exists to make it visible — and to do the first pass of it for you.
Nothing happens without you. AetherPilot proposes. You see a checklist — the task, the client and project it matched, the deadline, and the reason — and you check the boxes you want. Unchecked items are discarded. It never sends email, never marks your mail as read, and never writes to your data on its own.
Code-first (the recommended setting, in "Merge" mode). For the routine, well-understood items — e-filing notices, client intake, vendor reports — AetherTime works out the action in code. That is fast, free, deterministic, and identical on any model. A Code only mode runs with no AI whatsoever. The factory default is model-first, so this is a setting you turn on — and we recommend it.
AI-assisted. For everything else, a model reads the message and proposes an action. This is off by default and entirely optional. When you enable it, you choose the model:
The classification, matter matching, and word-count math run locally in every case. Turning AI on changes how the harder messages are interpreted; it does not move your data anywhere by default.
AetherTime grew up in insurance-defense litigation practice, so the workflow support is deepest for legal work: e-filing, records requests, matter intake, and deadline rules built around how a small firm actually runs. If you are a solo attorney or a small firm, that is the whole point.
Freelancers and consultants get the same engine without the legal vocabulary: email triage, a client-and-project hierarchy, and the admin-time measurement work for any practice where you read, sort, and file before you can bill. We are honest that the legal workflow is the most developed part of the product — it is where the real-world feedback comes from — and we are building the same depth for other practices next.
Your email contains client names, matters, and confidential detail. Sending that to a service is a decision you should get to make explicitly, per practice, not a condition of using a time tracker. AetherTime's answer is that the core work happens on your machine, the AI is opt-in, and you can point it at hardware you own.
The savings numbers are estimates, and we say so. Billable time is measured exactly. Admin time is derived from word counts and fixed baselines, and it improves as the app learns your work. A well-grounded estimate you can trust is worth more than a precise number that is quietly wrong.
Not by default. Classification, matter matching, and the word-count math run locally. The AI-assisted features are off unless you enable them, and you can point them at a local model on your own network. If you choose a cloud provider, it uses your own API key, and AetherTime operates no service of its own.
No. AetherPilot produces recommendations in an approval panel: each is a line with a checkbox, the matter it matched, any deadline, and the reason. Nothing is created until you apply the checked items. It never sends email and never marks your mail as read.
No. There is a code-first mode in which AetherTime works out the routine items — filing notices, client intake, vendor reports — deterministically, with no model involved, and the admin-time backfill never uses AI. AI-assisted triage is an optional enhancement.
It uses the word count of each document plus a fixed overhead, not a guess by a model, and it distinguishes message types with different reading speeds. The numbers are estimates and improve as the app learns your work, which is stated plainly in the time-savings documentation.
By matching the identifiers in the message — client and project numbers, claim numbers, names — against your own data, in code, with a defined priority order. You control the matching with a watch list of senders, phrases, and matter names, and you can reclassify history after changing it.
Yes, for the parts that are not legal-specific: email triage, client and project matching, and measuring the administrative time that never gets billed. The legal workflow (e-filing, records requests, statutory deadlines) is the most developed part of the product today, and we are building the same depth for other practices.
Private beta is open. Email us for a free copy before launch — no account, no credit card, no data sent anywhere.
Request Beta AccessAlso compare: AetherTime vs Clio · AetherTime vs Toggl · AetherTime vs Clockify · AetherTime vs Harvest · AetherTime vs spreadsheets · One-time purchase · Runs from a USB drive · For freelancers & consultants · How tuning works