Voice to Calendar
Say it the way you'd say it out loud. Composed handles the rest.
Composed is a voice-first planner for iPhone. Speak a sentence like "Dentist Tuesday at 2pm in Shibuya," and Composed uses Apple's on-device Speech framework to transcribe in real time, then builds the calendar event — with the place attached, prep tasks generated, and a leave-by time calculated. Syncs both directions with Apple Calendar and Google Calendar. Five events free.
The definition
Voice-to-calendar is the act of speaking a full natural-language event — date, time, place, and intent in one breath — and having an app parse the sentence, build a structured calendar entry, and write it to your iPhone's calendar without typing or follow-up prompts. Composed handles this with Apple's on-device Speech framework for transcription and a language model for event extraction, then syncs the result to Apple Calendar and Google Calendar.
That sentence does a lot of work. The "natural-language" part means you do not need to learn a command vocabulary — you say "Coffee with Maya at Stumptown Friday at 10" the same way you would say it to your partner. The "one breath" part means the parser handles date, time, place, recurrence, and implied tasks as a single utterance instead of asking three follow-up questions. The "structured entry" part means the result is a real calendar event with a real title, real time, real location, and real reminders — not a freeform note you have to clean up later.
Dictation
Apple Calendar and most stock apps treat voice as a typing substitute. You speak; words appear. You still have to tap fields, pick a time, set a date, choose a calendar.
Voice commands
Siri and Google Assistant respond to structured commands. They work well for short, unambiguous requests and tend to ask follow-up questions when the sentence carries place, recurrence, or partial information.
Voice-to-event
Composed takes the whole sentence — date, time, place, intent, recurrence — and writes a structured event in one shot. The same input produces the calendar entry, the prep checklist, the place, and the leave-by time.
What you can say
Composed handles the everyday — errands, meetings, friends, family, travel, recurring routines, multi-day stays. Each row below shows exactly what Composed extracts from the spoken sentence on the left.
You say
“Dentist Tuesday at 2pm in Shibuya.”
Event
Dentist appointment
Tuesday
Time
2:00 pm
Shibuya
Detail
Prep: bring insurance card. Leave by 1:32 pm.
You say
“Flight to LAX Friday morning, arrive at the gate by 8.”
Event
Flight to LAX
Friday
Time
8:00 am gate arrival
LAX
Detail
Departure tracking with airport buffer.
You say
“Pick up Sarah from JFK at 6:30 next Wednesday.”
Event
Pickup — Sarah
next Wednesday
Time
6:30 pm
JFK
Detail
Real-time travel from your location.
You say
“Coffee with Maya at Stumptown Friday at 10.”
Event
Coffee with Maya
Friday
Time
10:00 am
Stumptown Coffee
Detail
Place matched from your area, address attached.
You say
“Mom's birthday dinner next Saturday at 7 in the West Village.”
Event
Birthday dinner — Mom
next Saturday
Time
7:00 pm
West Village
Detail
Place discovery surfaces a calm shortlist.
You say
“Standup every Tuesday and Thursday at 9.”
Event
Standup
Tue + Thu, recurring
Time
9:00 am
Detail
Recurrence detected from "every".
You say
“Annual review with Priya April 8th at 11.”
Event
Annual review — Priya
April 8
Time
11:00 am
Detail
Prep: review prior quarter notes.
You say
“Pick up the dry cleaning before 6 today.”
Event
Pick up dry cleaning
today
Time
before 6:00 pm
Detail
Filed as a floating task, no clock pressure.
You say
“Soccer practice for Jacob Mondays at 5:30 at the Lincoln Field.”
Event
Jacob — soccer practice
Mondays, recurring
Time
5:30 pm
Lincoln Field
Detail
Departure alert factors in afternoon traffic.
You say
“Therapist Wednesday at noon.”
Event
Therapy
Wednesday
Time
12:00 pm
Detail
Quiet reminder ten minutes before.
You say
“Hotel check-in Friday, check-out Sunday at the Ace in New Orleans.”
Event
Hotel — Ace, New Orleans
Fri → Sun
Time
multi-day stay
New Orleans
Detail
Spans both days on your timeline.
You say
“Drop the car at the shop tomorrow morning, pick it up by 5.”
Event
Car service
tomorrow
Time
morning drop, 5:00 pm pickup
Detail
Two related tasks linked under one event.
You say
“Parent-teacher conference Thursday the 14th at 3:15 in room 204.”
Event
Parent-teacher conference
Thu Mar 14
Time
3:15 pm
Room 204
Detail
Prep: questions for the teacher.
You say
“Wedding rehearsal at 5 on Friday, dinner at 7 at Lilia.”
Event
Rehearsal + dinner
Friday
Time
5:00 pm, 7:00 pm
Lilia
Detail
Two linked events, single utterance.
You say
“Take meds tonight at 10.”
Event
Take medication
tonight
Time
10:00 pm
Detail
Calm reminder, no red badge, no guilt.
Composed treats partial input as partial. If you only said a date, the time stays open instead of being filled in with a guess.
Where it fits
Every voice-to-event tool on iPhone solves a slightly different problem. The table below maps what each one actually does today, so you can pick the one that matches how you plan. The short version: Siri and Google Assistant are great general-purpose voice assistants but lean on structured commands, Apple Calendar dictation is a faster keyboard but not a parser, and Composed is built specifically to take one natural-language sentence and turn it into a complete, ready-to-leave event.
| Capability | Composed | Siri | Google Assistant | Apple Calendar dictation |
|---|---|---|---|---|
| Open without a wake word | tap-to-talk | requires "Hey Siri" or hold | requires "Hey Google" or button | tap mic |
| Full-sentence natural language | limited; often asks follow-ups | limited; structured commands work best | transcribes only, no parsing | |
| Extract a specific location | with place discovery | if you say the exact address | if you say the exact name | |
| Generate AI prep tasks | ||||
| Calculate when to leave | with live traffic | basic, when location is set | basic, when location is set | |
| On-device speech recognition | Apple Speech | cloud-first | ||
| Reads + writes Apple Calendar | two-way | writes only | via Google Calendar sync | |
| Free for everyday use | 5 events free |
Why voice-first
The reason voice-to-calendar matters is the same reason planning fails in the first place — the right moment to add the event is rarely the right moment to open the app, tap the date picker, choose the calendar, set the time, and pick a category. The moment is somewhere between the doctor's office and the parking lot, or in the kitchen with a kid on each hip, or walking out the door with the dog leash in one hand and a coffee in the other. Voice closes the gap between thinking of the plan and capturing it.
In the car
Heard a soccer practice change in the carpool line. Speak it before you forget. Composed captures the new time, attaches the field, and recalculates the leave-by reminder for next Monday's practice.
Walking out the door
"Pick up the dry cleaning before 6." Composed files it as a floating task with a calm reminder, no time anxiety, no red badge — just a quiet check-in mid-afternoon while you have a moment.
After a phone call
Doctor's office calls to confirm Thursday at 9:15. Hang up, hold the microphone, say it out loud. The appointment, the time, and the address from your last visit all land on the calendar before you finish your next sip of coffee.
Living with ADHD
The hardest part of planning with ADHD is the gap between knowing and capturing. Voice cuts the gap to one sentence. Composed's three-layer reminders pick up from there — aware, ready, time to go — without overdue labels or red badges that make a bad day worse.
Composed was built for the moments planning actually happens — the in-between minutes, not the dedicated planning session that rarely arrives. The bar for voice-to-calendar is that you say the thing once, the way you'd say it to a friend, and it's handled. If you have to repeat yourself, simplify the sentence, or rephrase to match a command syntax, the tool has failed the moment. That bar is what Composed is built around.
Who it's for
Voice-to-calendar is not a feature for everyone — it's the right answer for a specific kind of planner. If you recognize yourself in any of the descriptions below, Composed is built for the way you already think.
Pediatrician changes, parent-teacher conferences, soccer practices, dance recitals, picture day. Family schedules generate more events per week than most calendars are designed to absorb. Voice lets you capture them in the moment they're announced — in the parking lot, on the school steps, mid-cooking — instead of writing them on a Post-it that gets lost in a coat pocket.
Flights, hotel check-ins, meeting blocks across time zones, dinners in a new city. Voice carries the date, time, and place into Composed in one breath. Departure tracking handles the airport buffer for flights and the live-traffic drive for everything else, so you stop running the math in your head every morning.
The combination of voice input, AI prep tasks, and a calm three-layer reminder system is built specifically for working-memory aid. The shortest path from "I need to remember this" to "it's captured" is a one-sentence voice memo. The shortest path from "I need to be ready" to "I'm leaving on time" is a leave-by alert that does the travel math for you.
Typing a full event into Apple Calendar — title, date, time, location, alert — takes between thirty and ninety seconds depending on how many fields you tab through. Speaking the same event takes about four. The math is unambiguous. Voice-to-calendar exists because the original calendar UX is faster as a sentence than as a form.
How it works
Voice-to-calendar in Composed runs as a five-stage pipeline: Apple's on-device Speech framework transcribes the audio in real time, a Claude-based parser extracts the event from the sentence, the result is structured into a calendar entry, an optional place is matched from nearby venues, and a leave-by time is calculated from live travel data. The whole sequence finishes in roughly the time it takes you to set the phone down.
You tap the microphone and speak. Apple's on-device Speech framework streams the transcription word-by-word to the screen so you can see what Composed is hearing as you talk. The audio stays on the iPhone — nothing is recorded, nothing is uploaded.
The transcript is parsed by a Claude-based extractor that pulls out the event title, date, time, location, recurrence, and any tasks implied by the sentence. Partial input stays partial. Ambiguous input is flagged so you can confirm rather than guessed.
If you mentioned a place by name or said something like "the dentist" or "dinner Saturday," Composed searches nearby venues and surfaces a short, calm list ranked by relevance and rating. You confirm with a tap, and the place is attached to the event with its address and hours.
Composed generates a short prep checklist based on what kind of event you described — bring the insurance card to the dentist, charge the camera before the trip, pick up a card before the birthday. You add or remove items with one tap.
For events with a location, Composed calculates real-time travel from where you are and sends a gentle "time to leave" reminder factoring in live traffic. For flights, the leave-by math includes the airport check-in and security buffer, so you do not arrive at the gate with thirty-one minutes to spare. Always confirm flight times with your airline.
Privacy
Composed never records or stores voice audio. On-device transcription means the audio stream stays on your iPhone. Only the resulting text is parsed, and your events are stored with row-level security so only you can read your own data. There is no advertising tracker, no resold profile, no audio archive.
Common questions
Yes. Composed is built around voice as the primary input — you say a plan in natural language, and Composed parses the date, time, location, and intent, then writes the event to its calendar (and to Apple Calendar or Google Calendar if you have sync enabled). You do not need to chain Siri shortcuts, dictate punctuation, or open a separate dictation field.
For natural-language voice-to-event capture, Composed is built specifically for the use case — it understands phrases like "Dentist Tuesday at 2pm in Shibuya" without needing trigger words, command syntax, or follow-up prompts. Apple Calendar dictation supports voice input but treats speech as raw text, not as a parsed event. Siri can create events but typically requires a structured back-and-forth and does not generate prep tasks or place suggestions.
Composed has its own dedicated voice button so you do not have to invoke Siri at all — you open the app, tap the microphone, and speak. Composed also reads from and writes to Apple Calendar, so any events Siri creates appear in Composed, and any events you speak into Composed appear in Apple Calendar and on every device synced to it.
Composed uses Apple's on-device speech recognition for real-time transcription, which means the audio stream stays on your iPhone. Composed does not record or store voice audio. When the optional cloud refinement pass is enabled, only the resulting text is sent for natural-language parsing — never the raw audio. Your events are protected by row-level security in our database, so only you can read your own data.
Yes. Composed handles relative dates ("next Wednesday", "in three weeks", "the 14th"), partial times ("morning", "before 6"), recurrence ("every Tuesday and Thursday"), multi-day spans ("check-in Friday, check-out Sunday"), and locations both specific ("Stumptown Coffee on Ave A") and approximate ("somewhere in the West Village"). It treats partial input as partial — if you only said a date, Composed leaves the time open instead of guessing.
How Composed listens, parses, and writes the event.
Automatic checklists for the events you say out loud.
Live travel time and a calm "time to leave" alert.
Three calm layers — aware, ready, time to go.
Voice input as a working-memory aid, no shame design.
The Siri-should-have-been vision behind voice-to-event.
Five events free. Calendar imports don't count toward the five. Speak the first one — Composed will handle the rest of the day.