Some conversations can never be repeated. A grandfather describing what the village looked like before the war, an aunt explaining how she emigrated alone at twenty, a witness recalling an event that now exists only in memory. These stories disappear if they aren't recorded in time, and they disappear a second time if they're recorded but never listened to again because no one has hours to transcribe them by hand.
This article isn't about work meetings or productivity. It's about something simpler and more important: turning a family recording into a written document you can read, print, share, and pass down to future generations, using AI transcription as the bridge between the audio and the memory itself.
Why it's worth recording now, not 'someday'
Oral history has an uncomfortable quirk: it only exists for as long as the person telling it is willing and able to tell it. You don't need a special occasion to record an elderly relative; a long lunch, a Sunday visit, or a video call with someone who lives far away is enough.
- Don't wait for the perfect setup: a phone's built-in microphone is enough to capture a voice clearly.
- Focus on concrete topics: a trade they practiced, a move, an era, a story that always gets told half-finished at family meals.
- Record in several short sessions instead of one two-hour marathon; memory works better and the person gets less tired.
What matters isn't the technical quality of the recording, but that it exists at all. Turning it into readable text can happen later, at your own pace.
Preparing the interview: questions that unlock memories
A good oral history interview isn't an interrogation, it's a guided conversation. A handful of open-ended questions work better than a rigid questionnaire:
- "What did a normal day look like when you were my age?"
- "What do you remember about the house where you grew up?"
- "Is there a decision in your life you'd change, or make exactly the same way?"
- "What would you want your grandchildren to know about that time?"
Leave room for silence, don't interrupt to correct dates or minor details, and let the person wander off-topic: often the most valuable memory shows up as a detour from the original question. If you're recording on video or audio from a phone, try to find a quiet space without background noise (TV, kitchen) — it makes for a much cleaner transcription afterward.
From recording to text: transcribing with AI
Once you have the audio or video of the interview, the next step is turning it into text. Doing this by hand — listening and typing sentence by sentence — can take several hours for every hour of recording; with an AI transcription tool like VOCAP, that same process takes just a few minutes.
The process is simple: you upload the audio or video file (an MP3, an M4A, a WhatsApp voice note in OPUS, an MP4 or MOV, up to 500 MB works fine) and within minutes you get the full transcript as plain text, plus an analysis with a summary, key points, dates mentioned, and standout quotes — especially useful when an interview covers several topics and you want to quickly find the part about, say, the grandparents' wedding or their arrival in a new city.
- The result can be copied or downloaded as Word (.doc) or .txt for further editing.
- The original audio is deleted from the servers once processing finishes, worth knowing if the recording contains sensitive family content.
- VOCAP supports over 50 languages, useful if the relative you're interviewing tells their story in a language other than English or mixes in regional expressions.
It's worth knowing that the tool doesn't automatically separate the interviewer's voice from the interviewee's, nor does it add timestamps, so if several people take part in the conversation you'll need to review the text and note who says what yourself. It also doesn't transcribe live: you record the conversation first, then upload the finished file. If you want to try it risk-free, VOCAP offers free minutes for a first transcription and one-time hour packs, no subscription required, whenever you need more.
Turning the transcript into a memoir
A raw transcript doesn't always read like a story: it has filler words, repetitions, and half-finished sentences, all natural parts of spoken language. The next step, done manually, is shaping that material into something readable:
- Organize the text into thematic blocks (childhood, work, family, historical moments lived through) rather than following the recording's chronological order.
- Keep the person's own expressions, sayings, and way of speaking when you write it down; that's part of what gives the testimony its value.
- Use the AI-generated summary and key points as an index to decide which fragments deserve to become chapters in a small family memoir.
- Add photos from the era mentioned in the interview alongside the matching text.
Many families end up laying this material out in a simple document, or even a printed booklet to hand out at the next family gathering or give as a gift on a special date.
Preserving and sharing the legacy
Having the text isn't the end of the process, it's the start of preserving it. A few practical recommendations:
- Keep the original audio file (if you decide to hold onto it yourself, outside the transcription tool) in at least two separate places, such as an external drive and a personal cloud account.
- Keep the transcribed text too, and if you've edited it, hold onto both versions: the literal one and the narrative one.
- Share the document with other family members; often someone else remembers a detail that fills in or adjusts the story.
- If there are several elderly relatives in the family, repeat the process with each of them: overlapping stories tend to reveal nuances that no single account can capture alone.
A family's oral history isn't a project you wrap up in a weekend, but every interview you transcribe is one more fragment that won't be lost.
FAQ
What audio format is best for recording an interview with an elderly relative?
Any phone voice recorder works fine; common formats like MP3, M4A, or WhatsApp voice notes (OPUS) all work well. The most important thing is to record in a quiet place with the phone close to the person speaking.
Can VOCAP identify who's speaking when several people take part in the interview?
No, VOCAP doesn't automatically separate voices by speaker. It delivers the full transcript as continuous text, so if you need to tell who said what, you'll need to review it manually.
Can I transcribe an interview recorded on video?
Yes, VOCAP accepts video files such as MP4 or MOV, as well as audio, with a limit of 500 MB per file.
What happens to the interview audio after it's transcribed?
The file is deleted from VOCAP's servers once processing finishes, which is handled through OpenAI and Anthropic. If you want to keep the original audio, save it yourself before or after uploading it.
Is the transcript ready to publish as a memoir right away?
The transcript delivers an accurate record of what was said, along with a summary and key points, but it usually needs manual editing afterward to turn it into a smooth narrative, since natural speech includes repetitions and unfinished sentences.