What to do in the 10 minutes after a job interview

You close the laptop. The adrenaline drops. You think that went okay, and you go make a coffee.

Two days later the recruiter asks how you'd feel about a follow-up round, and you realize you can't remember what you actually said about the budget question — or whether they asked about budgets at all. The interview is gone. What's left is a vague feeling and one embarrassing moment you'll replay for a week.

The ten minutes right after a call are the most valuable in your job search. Recall drops off a cliff fast, and everything useful — the questions, your answers, the things they let slip about the role — is only available in that window. Here's how to spend it.

Capture first, judge later

Don't evaluate how you did yet. You're a bad judge of your own performance while you're still nervous, and the analysis can wait. Get the raw material down while it exists:

The questions, in their words. "Tell me about a time you disagreed with your manager" is a different question from "how do you handle conflict." The phrasing tells you what they were probing for.
What you actually answered — not what you wish you'd said. Rough is fine.
Where you hesitated. The question that made you stall is the one to prepare before the next round.
What they pushed back on. A follow-up question is a signal your first answer didn't land.
What they told you about the role. Team size, why the position is open, what the first 90 days look like, who you'd report to. This is the part candidates forget fastest and need most.
What you still don't know. Every gap here is a question for the next stage.

How to get a transcript

Notes from memory work, and we'll come back to that. But a transcript is strictly better material, and getting one is easier than most candidates assume. Which option fits depends on your operating system, your budget, and whether you're comfortable with a notetaker bot visible in the call.

Two things before the list. Most of these you set up before a call — so if you're reading this straight out of one, skip to typing it out, then come back before the next round. And whichever you pick, you ask the interviewer first: recording law varies by country and by US state, and there's more on this below.

Prices, tiers and platform support are as of September 2026.

The transcript your call platform already makes

Google Meet, Zoom, and Teams all produce transcripts natively. The host usually controls whether it's on — so ask them to switch it on, or email afterwards and ask for a copy. Most will say yes; it's a normal request, and interviewers often want the record too.

This is the option to try first — there's nothing to install and no extra company holding your interviewer's voice. Two caveats. Availability depends on the host's plan and their admin's settings, so it isn't guaranteed on every account. And the transcript lands in the platform's own cloud: Drive for Meet, OneDrive or SharePoint for Teams, Zoom's cloud for Zoom.

The recording indicator makes the capture visible to everyone in the call, which is useful — but visibility is not agreement. You still ask.

Notetakers that join the call

These are hosted services that send a bot into the meeting to record and transcribe it. Your audio, transcript, and notes live on the vendor's servers rather than your machine. All three have genuine free tiers:

Otter — the free tier gives you 300 transcription minutes a month, but caps each single conversation at 30 minutes, which is shorter than most interviews; budget for the paid tier or expect the back half of the round to be missing. Pro is around $8.50 per user/month billed annually (about $17 monthly) with 1,200 recording minutes. Desktop, iOS and Android apps, plus a Chrome extension; works with Zoom, Teams, and Meet.
Fireflies — the widest platform coverage of the three: Zoom, Google Meet, Teams and 10+ others. There's a free tier with paid plans above it; their pricing page renders the current figures, and they move, so check before you commit. Worth noting if your interviews touch regulated work: it's SOC 2 Type II and GDPR compliant, AES-256 at rest, TLS in transit.
Fathom — the most generous free tier of the three: unlimited recordings and transcriptions, free forever. Premium is $20/month ($16 billed annually). Works with Google Meet, Zoom and Teams.

The trade-off is in the name. A bot joins the call, and the interviewer sees it in the participant list.

That cuts both ways. A visible bot is the most transparent option available — nobody can claim they weren't told. It's also the thing an interviewer is most likely to have an opinion about, in a conversation where you're being evaluated. Only you can judge which matters more for a given interview.

Recorders that don't join the call

These capture your computer's audio directly, so nothing appears in the participant list.

Be clear about what that does and doesn't change. It removes an awkward object from the call; it doesn't remove the need to ask — and it makes asking more important, because nothing else in the room will tell them. Recording someone who declined, or was never asked, is the version of this that can genuinely cost you the job, and in some places is illegal. Use these once you have a yes:

Granola — macOS, Windows, iOS and Android. The free tier gives you unlimited notes with history limited to 30 days; the paid Business plan is $14 per user/month and is what a solo job seeker buys, since there's no separate individual tier. It uses your computer's audio rather than inviting a bot. On where the data goes, read their security page rather than the marketing: Granola routes audio through third-party transcription providers (it names Deepgram and Assembly) and AI providers (it names OpenAI and Anthropic), and stores your transcripts and notes — not the audio — in a US-hosted AWS private cloud. It says it doesn't keep meeting audio after transcribing, and doesn't let those providers train on your data. Several companies you didn't choose still touch your interviewer's voice, which is worth knowing if you're in the EU.
OpenRec — open source under the MIT licence and macOS 15 or later only, so it's not an option on Windows or Linux. It captures system audio and your mic with no bot in the participant list, and produces speaker-labelled transcripts through AssemblyAI. The app costs nothing, but it runs on your AssemblyAI API key and that's metered pay-as-you-go — so "free" means free software plus a small per-hour transcription bill, not free transcription. Keys live in the macOS Keychain and recordings go to either OpenRec Cloud or your own Cloudflare R2 bucket. Audio does leave your machine — it reaches AssemblyAI and your storage bucket, and any AI provider you wire up for summaries — so the honest description is your keys and your storage, not offline.

Fathom also offers a bot-free capture mode, currently a Mac-only beta, which is where its Slack Huddle support lives. It's still the hosted service described above, so the audio goes to Fathom's servers either way — the bot is what disappears, not the upload.

Fully offline

Vibe is MIT-licensed and runs on macOS, Windows and Linux. It transcribes system audio and microphone input with speaker diarization, entirely on your device — its documentation describes fully offline transcription with no data leaving your machine. For transcription, it's the only option here where that's true, which makes it the pick if you want the interviewer's voice touching nothing but your own hard drive.

One caveat: that guarantee covers transcription. Vibe also offers summaries through the Claude API, which sends your text off the machine like anything else. Keep summaries off — or use its local Ollama option — and the offline claim holds.

Or just type it out

If none of the above fits — you're on a locked-down work laptop, the interviewer declined, the stakes don't justify the setup — open a blank document and type what you remember for ten minutes. Don't format it. Don't write in sentences if bullets are faster.

This is genuinely good enough. Notes are not a downgraded transcript; they're just a shorter one, and they're what most people will realistically have.

Ask before you record

Recording law varies by country and by US state — several require every party to consent, not just you — and in the EU, recording your interviewer means processing their personal data. Ask first. A short "do you mind if I record this so I can take proper notes?" costs you nothing.

If they say no, that's the end of it: type your notes instead. Plenty of companies have a blanket policy against candidates recording, so a refusal is usually the interviewer following a rule rather than telling you something about the role. And note that none of the tools above changes this answer — a recorder that stays out of the participant list is still recording someone who declined.

Turn it into verdicts

Raw material isn't the point. What you want is the answer to a harder question: which of my answers actually landed?

Go through what you captured question by question and sort each answer into one of three buckets — it landed, it missed, or it was built on a wrong assumption about the role. That third bucket is the one that changes outcomes. A technically fine answer aimed at the wrong seniority, the wrong team size, or the wrong priority reads as a miss to the interviewer and feels like a win to you, which is why it survives into the next round unexamined.

If your answers are rambling more than they're wrong, the fix is structural — the STAR method is the usual remedy.

This is also why speaker labels are worth caring about when you pick a tool. A transcript that marks who said what lets you separate their questions from your answers cleanly, so each verdict lands on the right pair. An unlabelled wall of text makes you reconstruct the boundaries by hand before you can judge anything — which is exactly the work you were trying to avoid.

This is the loop JobLane is built around: paste your transcript or your notes after each round, get a per-question read on what landed and what didn't, and have the next stage's prep start from what the last stage actually revealed instead of from scratch. Notes are enough — you don't need a recording to use it. That's the whole idea behind how JobLane works: each interview should make the next one sharper.

One last thing, while you're in the document: write down the questions you didn't get to ask. Those go in the follow-up email, and they're the cheapest way to look like the person who was paying attention.