Say it once.
It lands in the
right place.
Meridian is a voice-first task manager. Talk the way you talk, and it transcribes what you said, splits it into separate tasks, and files each one on the page it belongs to. Simple and seamless.
iPhone, Mac and web. $14.99 a month or $129 a year after the trial.
A recording of one sentence becoming three tasks, filed to Work, Shopping and Ideas.
Typing a task takes longer than having the thought.
So the thought waits. It waits in your head through the walk home, through the meeting, through the washing up, and by the time you are back at a keyboard it has either turned into three things or none. Every task app assumes you are already sitting down. Meridian assumes you are not.
Voice capture
Hold the button, say everything at once. Nothing to tap between thoughts, no fields to fill in afterwards.
Edit by voice
Say what changed. Move it to tomorrow, put it on Personal, mark it done, without touching the text.
Smart filing per page
One sentence can become three tasks on three different pages. Meridian decides where each one goes.
Works offline
Capture on the underground. The recording is held on device and files itself when you have signal again.
Cross-device sync
Said on the phone, there on the Mac before you sit down. Same list, same order, same state.
Per-page AI instructions
Tell each page what belongs on it, in your own words. Meridian reads that before it files anything.
You speak
Hold the mic on iPhone, or press the hotkey on Mac and speak over whatever you were doing.
It transcribes and splits
The transcript streams as you talk, then separates into individual tasks. Filler words are dropped.
It files them
Each task goes to the page whose instructions match it, with the routing shown so you can correct it in a tap.
One plan. Everything in it.
Seven days free, no card at the start. Cancel in one tap from the dashboard and keep access until the period ends.
7-day free trial, then $14.99/month. Cancel anytime.
Does it work if I speak quickly, or with an accent?
Yes. Transcription runs on the same models that handle dictation across dozens of English accents, and you can correct any word by voice or by typing. Corrections are not used to train anything.
What happens to my recordings?
Recordings are processed to extract the transcript and then deleted. They are never used for training models. The transcript is kept for 30 days and then removed unless you delete it sooner.
Can I use it without the AI filing?
Yes. Tasks land in your inbox if you prefer to file them yourself, or you can set per-page instructions that are deliberately broad so the AI files everything and you just correct the odd one.
Is there an Android app?
Not yet. Android is on the roadmap.
What if I cancel?
You keep full access until the end of the paid period. After that, existing tasks are still readable but capturing is paused until you resubscribe.