Une note d'entretien contient les antécédents professionnels d'une personne nommée, ses attentes salariales et votre lecture franche à son sujet. C’est exactement le genre de données personnelles que les gens oublient qu’ils téléchargent lorsqu’ils collent un débriefing dans un outil cloud. Ce guide montre comment capturer des notes d'entretien et des débriefings vocalement sur votre propre appareil, quelques minutes après la conversation, alors que votre impression est encore nette.
Read one of your own interview notes back and think about what is actually in it. A named person. Where they work now and what they earn. Why they want to leave. Something they told you about their manager that they would not put in writing. And underneath all of that, your candid, unedited assessment of them.
Then think about where that note went. For a lot of recruiters and hiring managers it went into a cloud notes tool or a chat window with an AI assistant, because that is where the tidying-up happens. A named person's salary and your opinion of them got uploaded to a third party so that a machine could fix the punctuation.
There is a version of this workflow where that never happens. You speak the debrief into your own device in the two minutes after the interview, on-device AI structures it into clean text, and it lands in a private searchable vault you can look back through when you compare five candidates a week later. Nothing about the candidate is uploaded to become a note.
One thing up front, said plainly: nothing here is legal or HR advice, and Yaps does not score, rank, or assess candidates for you.
It is easy to treat interview notes as informal. They feel like working memory, so they get handled far more casually than a CV or an offer letter.
Look at the content, though. A single debrief typically contains a person's full name, their current employer, their notice period, what they earn and what they are asking for, a reason for leaving that may involve a conflict at their current job, and your subjective judgement of their competence and character. Some of it was given to you in confidence because they are quietly looking. Most candidates never see a word of what you wrote about them.
That is personal data about an identified person, and in most jurisdictions it is treated as such. Your organisation almost certainly has obligations about where it is stored, how long it is kept, who can see it, and what happens if the candidate asks for a copy. Those obligations do not switch off because you wrote the note in a hurry.
So where does the note go when you tidy it up? If you dictate the debrief into a cloud transcription service, or paste it into a cloud assistant to be cleaned up, you have handed a named candidate's data to a vendor whose retention policy you did not write, on infrastructure you do not control, possibly in a country you did not choose. You may have a lawful basis and an agreement covering that. You may also have never checked.
Keeping the capture on your own device removes the question for the note itself. It does not remove your wider duties or replace your organisation's policy. It does mean the raw, candid version never travels.
Set privacy aside for a moment, because there is a second problem here.
Interview notes decay fast. You run four screens back to back, promise yourself you will write them up properly later, and later turns out to be the end of the day. By then the second and third candidates have blurred, one specific answer has attached itself to the wrong person, and the reason you liked somebody has flattened into "good communicator". That is not a record. That is a summary of a feeling.
The specifics are what you actually needed. Not "strong on stakeholder management" but "described the exact moment she told the finance director the deadline was not achievable, and what she offered instead". The first could describe half your pipeline. The second is evidence, and evidence is what makes a hiring debate short.
There is a fairness argument here too. Notes reconstructed at the end of a long day are more exposed to recency and halo effects than notes captured immediately after each conversation. The candidate you saw at 9am gets thinner coverage than the one you saw at 4pm, not because they were worse, but because you had four hours less decay. Writing evidence against your scorecard headings while the conversation is still in your head gives every candidate a comparable record, which is also what your organisation will want if a decision is ever questioned.
Without Yaps
You reconstruct four conversations from a page of half-legible bullets. Two candidates blur together, one good answer gets credited to the wrong person, and the notes come out as adjectives rather than evidence. A week later, when the panel disagrees, there is nothing concrete to point at.
With Yaps
You speak the debrief in the gap before the next call, while the specifics are still sharp. On-device cleanup turns the ramble into structured text, and it saves into a private vault you can search by name, role, or skill when the comparison happens later.
Yaps is a voice assistant that runs on your own device. For hiring work it does four concrete things, and all four happen locally.
It turns a spoken debrief into clean text. Press the Yaps hotkey (on desktop the key is Fn, and on Android it is the dictation button on the Yaps keyboard) and talk the way you would brief a colleague, restarts and asides included. Recognition runs on your device, so the audio of you talking about a candidate is never uploaded to become text. It works across every app, so you can speak straight into your notes app, an email to the hiring manager, or the comment box where your team records decisions. The dictation feature page covers the mechanics.
It cleans up the ramble on-device. Cleanup is on by default and runs locally. It strips filler words and self-corrections, fixes punctuation and capitalisation, and turns a spoken list into an actual list. So "so, um, on the stakeholder thing she was, sorry, she was really specific, she talked about pushing back on a finance director and then, right, offering a phased delivery instead" arrives as a clean sentence you would be happy for a hiring manager to read. You get the speed of talking with the finish of writing, and no internet round-trip in between.
It keeps everything in a private, searchable vault. Debriefs land in a local vault on your device, searchable without a cloud index of your candidate pipeline. Notes support checklists and kanban boards as well as plain text, so a role can be a board with a card per candidate, and everything exports to Markdown or plain text. The voice notes feature goes deeper on the structure.
It syncs between your phone and your laptop. Vault syncing is a premium feature that pairs your own devices with a QR code and moves notes over your local network or an encrypted peer-to-peer link, not through a central server. The debrief you speak walking back from a meeting room is on your laptop when you sit down.
One more piece needs a careful boundary. If you recorded the interview yourself, with the candidate's consent and in line with your policy, the Yaps Studio editor imports that audio file and transcribes it offline into text or an SRT subtitle file. That is a recording you already own becoming private text on hardware you control, not a bot sitting in your call. Yaps does not join meetings, record interviews automatically, or do live meeting transcription. For the longer argument on why that boundary is a feature, see taking notes without a meeting bot.
That single question sorts the options better than any feature list. Hold the on-device approach against the two things recruiters most commonly reach for: an AI notetaker that joins the call, and typing the debrief into a cloud document.
| Where the words go | Yaps (spoken debrief, on-device) | Cloud AI notetaker in the call | Typing into a cloud document |
|---|---|---|---|
| Audio sent to a vendor | None, it stays on your device | Yes, the whole interview | No audio |
| Where the note text rests | Local vault on your device | Vendor cloud account | Vendor cloud account |
| Works with no internet | Yes, for capture and search | No | Partly, then it syncs |
| Adds a third party to the interview | No, you speak afterwards | Yes, a bot in the call | No |
| Consent conversation required | Not for your own notes | Yes, recording consent | Not for your own notes |
| Scores or ranks the candidate | No, by design | Some do | No |
| Searchable when you compare later | Yes, local vault search | Yes, in the vendor cloud | Yes, in the vendor cloud |
| Writes into your ATS automatically | No, paste-ready text | Some integrate | No, paste it |
Read the last row as an honest trade rather than a defeat. Yaps has no integration with any applicant tracking system, so the output is clean, paste-ready text you copy into whatever field your ATS gives you. If automatic write-back is a hard requirement, a tool built into that platform is the right answer for that step, and the sensitive capture can still stay local.
The rows above it are where the on-device approach earns its place. A tool that never uploads a recording of your assessment cannot leak it, cannot retain it after you delete it, and cannot quietly change its terms next quarter.
Your candid read on a named candidate is the most sensitive thing you write all week. It does not need to leave your desk to become clean text.
Yaps
The loop below takes about two minutes per candidate and it replaces the evening write-up entirely.
Install Yaps, start the trial, and do one test dictation with the internet switched off so you have seen for yourself that nothing is being uploaded.
Stop scheduling back to back. The debrief is only accurate before the next conversation writes over it.
Name and role first, then evidence against each scorecard heading, then open questions and your recommendation.
Filler and false starts go, punctuation and lists arrive. No editing pass, no second tool, no upload.
The note saves to your private vault. Copy the clean text into your ATS or the decision record your team uses.
An honest scope is more useful than a long feature list, and in hiring the boundaries matter.
It does not record or join interviews. Yaps has no meeting bot. It does not sit in your video call, does not transcribe the conversation live, and does not capture the candidate speaking. This is a note-and-debrief workflow: you talk, afterwards, about what you observed. If you did record the interview yourself with the candidate's consent, Studio transcribes that audio file offline, which is a different thing from a bot in the room.
It does not score, rank, assess, or shortlist candidates. Yaps performs no automated decision-making about people. It does not rate an answer, compare candidates, produce a fit percentage, or recommend who to advance. That is deliberate: automated candidate assessment carries real legal and fairness risk, it is regulated in a growing number of places, and it is not a feature we are quietly adding to a notes app. Yaps writes down what you said. The judgement stays yours, and so does the accountability for it.
Recording an interview needs the candidate's consent, and the rules vary. Consent requirements differ by country and by state, and your organisation may have its own rules on top. If you plan to record, tell the candidate, get their agreement, and follow your policy. Using Yaps to transcribe a file you recorded does not change any of that.
There is no integration with any applicant tracking system. No connector, no automatic write-back, no sync. Yaps gives you clean, structured, paste-ready text and you copy it into your ATS. If automatic write-back is a hard requirement, use a tool built into that platform for that step.
There is no iPhone or iPad app yet. Yaps ships on Android, Windows, macOS, and Linux today, plus a Chrome extension for saving web pages into your vault. iOS is coming soon. If your recruiting team lives entirely on iPhones, this workflow is not available to you yet.
Read-aloud voices are English speakers in practice. Yaps reads text back to you with 18 voices on desktop and 2 on mobile, which suits reviewing a long note while walking. If you need read-aloud across many languages, a multilingual cloud reader is the better fit. Dictation itself is multilingual, at about 25 auto-detected languages.
The debrief habit pays off when you sit down to make a shortlist, usually days after the conversations happened.
Open your vault and search. Because the notes are text rather than recordings, and search runs locally, you can pull up every candidate who mentioned a specific tool, every debrief for a role, or the one person who described a migration in enough detail that you believed them. There is no cloud index of your pipeline and no vendor holding a list of who you interviewed for what.
Structure helps here more than people expect. Run a role as a kanban board with a card per candidate, moved from screened to panel to offer, and the state of the search is a screen rather than a memory. Run a panel round as a checklist and you have a record of what was asked, not only what stood out. See turning voice notes into action items for the habit.
When the decision is made, the notes go where your organisation requires and are kept or deleted under your own retention policy. That part is your responsibility, not a feature, though a local vault does make deletion simpler because there is no vendor copy to chase.
Start with Yaps for the capture layer of hiring. The interview note is the most sensitive document in the process and the one place where a cloud round-trip buys you nothing. Speak the debrief in the two minutes after the interview, let on-device cleanup turn it into structured text, and keep it in a private vault you can search when the comparison happens. The notes get more accurate and more comparable at the same time as they get more private, which is a rare combination.
Two places where something else wins. If your process depends on notes writing themselves into a specific applicant tracking system, use a tool built into that platform for that step and keep the candid version local. If your team is entirely on iPhone and iPad, wait for the iOS release.
For everyone else, the change is small and it starts on Monday. Stop writing four debriefs from memory at 6pm. Speak each one while you still remember what the candidate actually said, on a device that keeps it to itself. For the wider argument about voice input and privacy, read is voice typing private, and offline dictation covers what still works when the connection does not.
Yaps is the strongest default, because dictation, cleanup, storage, and search all run on your own device, so a named candidate's details and your candid assessment never travel to a vendor to become clean text. It runs on Android, Windows, macOS, and Linux, and works offline. If you need automatic write-back into a specific applicant tracking system, use a tool built into that platform for that step.
Yes, and that is the workflow described here. Instead of recording the interview, you speak your own debrief straight after it, and on-device AI turns your speech into clean, structured text. No bot joins the call and the only voice involved is yours, so you are writing down your observations rather than capturing a conversation.
It depends on where you are, so treat this as a prompt to check rather than an answer. Recording a conversation requires the other person's consent in many countries and in a number of United States states, and your own organisation may impose stricter requirements on top. If you intend to record, tell the candidate, get their agreement, and follow your policy. Nothing here is legal advice.
No. Yaps does not join meetings, appear in your video call, record interviews automatically, or transcribe conversations live. It is a note-and-debrief tool: you speak your own summary afterwards and it becomes clean text on your device. If you recorded an interview yourself with consent, the Yaps Studio editor transcribes that audio file offline.
No, and that is deliberate. Yaps performs no automated decision-making about people: it does not rate answers, produce fit scores, compare candidates, or recommend who to advance. Automated candidate assessment carries real legal and fairness risk, so Yaps stays on the writing side. The judgement and the accountability remain yours.
Capture them immediately and against a fixed structure. Block two minutes after every interview, speak the candidate name and role first, then concrete evidence under each scorecard heading, then your open questions and recommendation. Quote what the candidate actually said rather than writing adjectives about them. Notes captured this way are far more comparable than anything reconstructed at the end of a hiring day.
No. There is no integration, connector, or automatic write-back to any applicant tracking system, CRM, or HR platform. Yaps produces clean, structured, paste-ready text that you copy into whatever field your ATS provides, and notes export to Markdown and plain text. If write-back is a hard requirement, use a tool built into that platform for that step.
Yes, if you already hold the audio file and recorded it lawfully with the candidate's consent. The Yaps Studio editor imports the file and transcribes it offline into text or an SRT subtitle file, with nothing uploaded. It works only on files you already have, because Yaps does not join or record live calls.
No. Notes are stored locally on your own device in a searchable vault, and the audio of your dictation is not uploaded to be turned into text. If you enable premium sync, notes move between your own phone and laptop over your local network or an encrypted peer-to-peer link paired with a QR code, not through a central server. Deletion is therefore something you control directly.
Yaps offers a 7-day free trial, and after that it is a paid subscription: Pro at $15 a month and Max at $25 a month. There is no free tier. An account is required only for billing, not for the dictation, cleanup, and local search that run on your device.
Not yet. Yaps ships on Android, Windows, macOS, and Linux today, plus a Chrome extension for saving web pages into your vault, and the iPhone and iPad app is coming soon. If your recruiting team works entirely on Apple mobile devices, this workflow is not available to you yet.
Yes for the notes. Dictation recognises about 25 spoken languages and detects them automatically from your speech, so a debrief spoken in German, Spanish, French, or English needs no setting change. Read-aloud voices are a separate matter and are English speakers in practice, so a multilingual cloud reader suits that job better.
Search runs locally in your vault, so you can look up a candidate name, a role, a tool they mentioned, or a phrase from your own debrief without querying a server. The trick is consistency: open every debrief with the role and the candidate name, and keep your scorecard headings in the same order. Running a role as a kanban board with one card per candidate also turns the state of the search into a screen rather than a memory.