Voice journal: how to keep a diary by speaking, and who can hear it
Some people think best with a pen. Many more think best out loud. If you have ever started a diary and given up because writing felt like work, try saying it instead. A voice entry takes a couple of minutes, you can make it on a walk or in the car after you've parked, and it keeps something text never will: how you sounded that day.
Play back a recording from ten years ago and you hear the tiredness or the excitement in your own voice, the laugh in the middle of a sentence, the pause before the hard part. That is why a voice note is often the most vivid part of a memory.
How to make a voice entry worth hearing again
A rambling twenty-minute recording is hard to listen to later. A few habits help:
- One moment per recording. Not "my week", but "the conversation with Dad on Sunday".
- Start with the frame: where you were, when, who was there. Those details fade first, and they are what helps you find your way back into a memory later.
- Say what happened, then what you felt. In that order. The facts give the feeling a place to stand.
- Use your senses. What you saw, heard, smelled. "The kitchen smelled of burnt toast" brings back more than "it was morning".
- End with why it matters, or what you think will happen next. Future you will want to know.
- Keep it short. Two to five minutes is plenty. Don't stop to fix your sentences; the stumbles are part of it.
When you don't know what to say
- What happened today that you'd want to remember in ten years?
- Tell the story of the day you met someone important to you.
- Describe the place where you grew up, room by room.
- What is worrying you right now? Say it out loud, then say what you'd tell a friend in your place.
- What did you learn this year that you didn't know last year?
- Describe a photo from your gallery as if to someone who can't see it.
The privacy question: where does your voice go?
"Voice journal" can mean three different technologies, and each sends your voice somewhere different. Most apps don't spell this out, so it is worth knowing what to ask.
1. A recording
The app saves an audio file. On your phone, it is only as private as your phone. If the app syncs to the cloud, the file sits on its servers too. Ask whether recordings are encrypted with your own key or stored as ordinary files, and whether they are included when you export your data.
2. Live dictation
You speak and the words appear as text. Most apps don't build their own recognizer: they use the speech recognition that Apple or Google provide on the phone. That audio may be processed on the device or on Apple's or Google's servers. An app can ask for recognition on the device only, where the phone supports it for your language; ask whether yours does. The microphone key on your keyboard is a separate system feature with its own rules: on an iPhone, Apple's Siri, Dictation & Privacy page explains that Keyboard Settings show whether dictation is processed on your device. On Android, the developer documentation says that unless an app asks for offline recognition, "either network or offline recognition engines may be used".
3. AI transcription of a recording
You record first, then ask the app to turn the recording into text. Usually the audio is sent to an AI service that runs on a server. Ask whether that is optional, which company receives the audio, and whether it is kept afterwards.
And one more practical point: anyone who can open your journal can play your recordings. If you share a tablet with your family, a voice diary on it is less private than a paper notebook in a drawer.
Voice in Lifeline
In Lifeline a voice note belongs to a memory: you tell the story of that moment, and the recording sits next to its photos, the song that was playing and how it felt. Here is exactly how it works:
- Recording. The free version keeps one voice note per memory; Premium has no limit. Recordings are saved on your phone and synced to your account in Google's cloud. They are not encrypted with your master password, even in a locked memory; neither are their names. Only a memory's words are: the text, your reflections, the AI's questions and your answers.
- Live dictation. Lifeline uses the speech recognition Apple or Google provide on your phone, so the text appears as you speak. Lifeline itself doesn't send this audio anywhere. On Android it asks for recognition on the phone itself, where your phone has it for your language; otherwise Google's recognizer may use its servers. On iPhone, Apple's speech recognition may process the audio on Apple's servers.
- "To text". You can turn a recorded note (up to about ten minutes) into text that is added to the memory. This is free. It sends that one recording, through our server, to Google's Gemini model, only after you've agreed to AI features, and asks again if the memory is locked. Our server doesn't keep a copy of the audio after the text comes back. Under Google's paid-service terms, Google doesn't use it to improve its products, though it keeps logs for a limited time to prevent abuse and for legal reasons. The text it produces is encrypted like anything you type, if the memory is locked.
- Export. The free export includes every voice note, in a ZIP with your texts and photos.
Record the voices around you, too
The recording you may treasure most in twenty years isn't your own voice. It's your grandmother telling how she met your grandfather, your father explaining how to fix something, your child at four describing a dream. A voice is easy to lose and hard to imagine back, and no photo brings it back. Ask someone you love to tell you one story and press record. (Some questions to ask your parents to get you started.)