Journaling
How do you record your dreams?
You record a dream by speaking or writing your memory of it as soon as you wake, because no device can capture the dream itself (Horikawa et al., 2013). Speaking tends to keep more detail, while writing may make it easier to spot links with your day (Casagrande & Cortini, 2008; Schredl et al., 2019). Pick whichever you can manage half asleep, and if you speak, turn the recording into text later.
Last updated: 2026-10-13.
Can you record a dream and watch it later?
No device can record a dream for you to watch later; the closest research decoded broad categories of what people saw as they drifted off, inside an MRI scanner (Horikawa et al., 2013). In that study, three people were woken again and again from light sleep soon after they fell asleep (ATR Brain Information Communication Research Laboratory Group [ATR], n.d.). Each was woken until they had given at least 200 reports of what they had just seen.
The researchers first trained decoders on each person’s brain activity while they looked at pictures. From the activity just before each waking, the decoders could then predict which kinds of object the person went on to report, out of about 20 categories (ATR, n.d.; Horikawa et al., 2013). That is a long way from a video of a dream, and headlines about machines that “record and replay” dreams go well beyond it. For now, recording a dream means recording your report of it.
Is it better to speak a dream or write it down?
Speaking and writing each have strengths, and we found only two studies that compare them (Schredl et al., 2019). In the first, 302 adults aged 18 to 40 kept a dream diary for 14 days, tape-recording each dream and also writing it down (Casagrande & Cortini, 2008). The written versions lost some of the sensory detail of the spoken ones. A later paper points out that everyone reported the same dream twice, so motivation may have shaped those results (Schredl et al., 2019).
The second study, reported in the same paper, was smaller. Of 45 people assigned at random to speak or to write their dreams for a week, 40 were analyzed. Spoken reports were three times as long, about 310 words against 101 on average. But the writers recorded more dreams over the week (2.6 against 1.7 on average), and they found more links per dream between what they dreamed and their waking life. Two studies, one of them small, can’t settle the question, so treat these as leads.
| What matters | Speaking | Writing | Both |
|---|---|---|---|
| Detail kept | Longer reports, more sensory detail | Shorter; some detail lost | The recording keeps the detail |
| In the dark | No light needed; eyes can stay shut | Needs a small light | Speak now, write once you’re up |
| Rereading and searching | Slow to replay a long recording | Quick to scan and search | Text to scan, audio to go back to |
| Links to your day | Fewer found in one study | More found in the same study | Look for links in the text |
| Someone asleep beside you | May wake them; whisper | Quiet | Whisper a few words, write later |
You can also do both: speak while you’re groggy, then type a few lines later, so you have text to search when you look for patterns in your dreams.
Why can speaking be easier when you’re half asleep?
Speaking can be easier because the first minutes after waking are foggy (Hilditch & McHill, 2019). Sleep inertia, the grogginess you feel on waking, is linked with significant drops in mental performance that fade the longer you’re awake. The fog starts to lift quickly, though full recovery may take an hour or more.
Speaking asks less of you in that state: no light, no spelling, and your eyes can stay closed. That is our reasoning rather than a finding, since we found no study that tested speaking against writing or typing in the first groggy minutes.
What should you ask before trusting an app with your voice?
A dream journal holds personal material, so put these questions to any voice app, or to an AI that reads dreams:
- Where is speech turned into text: on the phone, or on a company’s server?
- Is the audio kept, or uploaded anywhere?
- Do you need an account?
- Where is the journal stored?
- Can you export everything, and delete everything?
- Are your words used to train AI models?
- Can you turn analytics off?
Then read the listing. Each App Store product page shows some of the data types an app may collect, such as Audio Data, meaning your voice or sound recordings (Apple, n.d.-a). It also shows whether that data is linked to you or used to track you. The developer supplies these answers and is responsible for keeping them accurate, and every listing must link a privacy policy. Read the answers as the developer’s own statement.
When an iPhone supports on-device speech recognition, an app can require it, so the audio stays on the device and none is sent over the network (Apple, n.d.-b). Apple notes that on-device recognition won’t be as accurate.
How can you record dreams without an app?
You don’t need a dream journal app: the phone and paper you already have will do. Your phone’s built-in voice recorder works, or dictation into a notes app, followed by one written line later in the day. So does a notebook with a small light, or a single keyword on a sticky note to jog your memory at breakfast. If someone sleeps beside you, whisper or write. Keep whatever you use within reach of the bed, the first habit in the 7-night dream recall plan. Say or write the same few things each time; the fields in the dream journal template work either way.
More questions about recording dreams
Should you transcribe voice notes later?
Transcribing helps if you want to look back: in a small study, people who wrote their dreams found more links with their waking life than people who spoke them (Schredl et al., 2019). The authors suggest that researchers ask people to listen to a recording at least once, or turn it into text, before looking for links. For your own journal, a few typed lines are enough.
Is there a gadget that records dreams?
No. The research gets nowhere near a recording: decoders built from MRI brain scans could only predict broad categories of what people reported seeing as they fell asleep (Horikawa et al., 2013).
Can you record a dream in the middle of the night?
If you wake on your own, a few whispered or scribbled words are enough to note it. You don’t need an alarm for this; the 7-night dream recall plan saves it for nights you happen to wake.
Does it matter if you ramble?
No. Spoken reports run long anyway: in one small study they were three times the length of written ones (Schredl et al., 2019). Say everything, then write a short summary later if you want one.
Speech that stays on your phone
In Dreams Come True, a voice-first dream journal app for iPhone, iOS transcribes your speech on the phone, the audio is never uploaded, and there is no account to create. The privacy policy covers the rest.
Sources
- Apple. (n.d.-a). App privacy details on the App Store. Apple Developer. developer.apple.com/app-store/app-privacy-details/
- Apple. (n.d.-b). requiresOnDeviceRecognition. Apple Developer Documentation. developer.apple.com/documentation/speech/sfspeechrecognitionrequest/requiresondevicerecognition
- ATR Brain Information Communication Research Laboratory Group. (n.d.). Neural decoding of visual imagery during sleep. bicr.atr.jp/dni/en/research-projects/neural-decoding-of-visual-imagery-during-sleep/
- Casagrande, M., & Cortini, P. (2008). Spoken and written dream communication: Differences and methodological aspects. Consciousness and Cognition, 17(1), 145–158. doi.org/10.1016/j.concog.2007.11.002
- Hilditch, C. J., & McHill, A. W. (2019). Sleep inertia: Current insights. Nature and Science of Sleep, 11, 155–165. doi.org/10.2147/NSS.S188911
- Horikawa, T., Tamaki, M., Miyawaki, Y., & Kamitani, Y. (2013). Neural decoding of visual imagery during sleep. Science, 340(6132), 639–642. doi.org/10.1126/science.1234330
- Schredl, M., Dreer, J., Mösle, A., Rall, M., Rauch, L., Rose, S., & Seuffert, S. (2019). Voice-recorded vs. written dream reports: A research note. International Journal of Dream Research, 12(1), 138–140. doi.org/10.11588/ijodr.2019.1.58801