Why we built this
The problem with task managers isn’t that they’re missing features. It’s that they’re all built around the same assumption: the user knows exactly which list this belongs on, what priority it deserves, and when it’s due — and is happy to spend three seconds filling fields to say so.
Most of the time, that’s wrong. You already said all of that in the sentence “Email Alex about the deck Friday.” You’re just being asked to retype it into form fields because the software can’t parse plain English.
The premise
EchoTime is the experiment of treating that retyping as a bug. You type the sentence. We pull out the action, the person, the deadline, the project. You hit enter and you’re done. If we got something wrong, edit it in place — but we’d rather be right ~85% of the time and let you fix the rest than make you do data entry on 100% of inputs.
Why now
A few things converged:
- Gemini’s extraction quality crossed the threshold where it’s reliable enough to be the default path, not a power-user feature.
- Firestore’s live snapshots are mature and cheap. Real-time sync across every device used to be infrastructure work; now it’s a library call.
- The web caught up to native. View Transitions, the
prefers-reduced-motionmedia query, optimistic mutations as a pattern — there’s no longer a 100ms tax for choosing the browser over a native app shell.
The dashboard you’ll find at app.echotime.ai is what happens when you take those three together and stop apologizing for the result.
What’s next
We’re shipping the public marketing site now (you’re reading it), the public Echo pages soon, and native iOS/Android/macOS apps after the web version stabilizes. The blog is where we’ll talk about the engineering as we go.