Transcribe and Translate
Candidate Interviews

MirrorCaption transcribes and translates candidate interviews in real time — streaming word-by-word in any of 50+ selectable languages, directly in your browser, with no bot visible to the candidate. Unlike post-meeting tools such as Metaview or Sonix, it shows you what the candidate said while they are still talking, so you can ask a follow-up before the moment passes.

Every recruiter running cross-language interviews faces the same problem: a candidate answers in their native language, you catch roughly what they meant, and by the time you have processed it, the opportunity to probe has gone. A transcript that arrives after the call helps with documentation. It does not help with the next 40 minutes of interview. Real-time translation for distributed teams solves a similar problem — the recruiting context just has higher stakes per conversation.

Remote's 2025 Global Workforce Report found that 73% of HR leaders expect more than half of new hires to be international by 2026. The interviewing workflow hasn't kept pace. Most teams rely on post-meeting tools built for English-primary, synchronous calls — or hire bilingual recruiters for every language market, which does not scale. MirrorCaption closes that gap with a browser tab that streams the transcript as the conversation happens.

Why Real-Time Translation Changes the Interview

When the hiring decision happens during the interview, post-meeting tools are structurally too late. If you miss a critical answer at minute 12, no amount of transcript polish after the call restores the chance to follow up.

Real-time translation is not a speed feature in this context. It is the difference between evaluating the candidate and evaluating your comprehension of the candidate. Forcing candidates to interview in a non-native language compounds the problem: research consistently finds that non-native speakers are rated lower in unilingual interview settings than their actual competence warrants, independent of qualification.

Here are four recruiting scenarios where this matters most:

How MirrorCaption Works During a Job Interview

MirrorCaption runs in a browser tab alongside your video call. There is nothing to configure for the candidate and nothing to install on either side.

For video interviews (Meet mode)

1. Open MirrorCaption in desktop Chrome or Microsoft Edge alongside your browser-based Zoom, Google Meet, Microsoft Teams, or Webex call
2. Select the meeting tab when prompted to share tab audio (the same browser permission used to share a screen in a video call)
3. Choose source and target languages — the language the candidate speaks and the language you want to read
4. Start the session — the live transcript streams word-by-word, with automatic speaker labels (Interviewer, Candidate) as each voice speaks
5. Export after the interview — full transcript in Markdown or plain text, with speaker labels, for documentation or sharing with the hiring team

For phone or in-person interviews (Talk mode)

Open MirrorCaption in Chrome on your mobile device and select Talk mode. Place the phone facing the speaker or on the table between you. The transcript and translation stream live on your screen. No video call is required. Talk mode works best in Chrome on mobile, so the candidate does not need to be on a video platform at all.

Speaker detection identifies distinct voices automatically. You can rename labels from "Speaker 1 / Speaker 2" to the actual names of participants after the session.

No Bot in the Meeting Room

Many meeting note tools work during live calls by joining as a bot participant. Fireflies can join through fred@fireflies.ai or calendar auto-join settings, while its Chrome extension offers a Google Meet path without the bot. OtterPilot appears in the participant list with its own name. Most meeting bots announce themselves when they join.

MirrorCaption captures audio from your own browser tab. It never appears in the meeting's participant list. The candidate sees a standard meeting interface with the usual participants only.

A note on disclosure: informing candidates that transcription is running is standard good practice and a legal requirement in many jurisdictions under GDPR and equivalent frameworks. MirrorCaption does not remove that responsibility. It removes the visual cue in the meeting room that changes how candidates present themselves — which is a separate problem.

Many enterprise IT security policies block third-party meeting bots from accessing corporate call infrastructure. Because MirrorCaption never joins the meeting itself, there is no admin approval needed for the tool. It runs as a web app in your browser tab, subject to the same browser and screen-capture permissions as any tab on your machine. Most recruiting teams can self-serve without an IT ticket. For guidance on screen-capture policies, check with your IT team — workplace policies vary.

Interview Candidates in Their Native Language

MirrorCaption supports 50+ selectable languages including Mandarin, Japanese, Korean, Arabic, Hindi, Spanish, French, German, Portuguese, Russian, Hebrew, Turkish, Polish, and more. Both the source language (what the candidate speaks) and the target language (what you read) are selected independently. The transcript shows the original alongside the translation simultaneously, not one replacing the other.

Some nuances matter specifically in hiring contexts:

When a Japanese candidate says 「それは、少し難しいかもしれません」 — literally "That might be a little difficult" — the surface translation is accurate. The cultural signal, that this is a considered objection rather than mild hesitation, requires the surrounding context to interpret. Having the original text alongside the translation lets you ask a clarifying question while you still can, not three days later reading a polished summary.

When a Spanish-speaking candidate says "Podría hacerlo, pero dependería de los recursos" ("I could do it, but it would depend on the resources"), the conditional framing is a signal worth probing. In a post-meeting transcript reviewed 90 minutes later, the prompt to ask is gone. With MirrorCaption streaming live, you catch it in the conversation.

This also has a fairness dimension. Conducting candidate interviews in English when it is not the candidate's native language introduces noise into the assessment that has nothing to do with the role. For teams focused on multilingual remote collaboration, the same principle that governs team meetings applies to the process that builds the team.

For a broader comparison of multilingual transcription tools across use cases, see the multilingual transcription guide.

Audio Privacy During Candidate Interviews

Candidate interview data is personal data under GDPR. Voice recordings are treated as biometric data, and mishandling interview recordings has resulted in substantial GDPR enforcement actions since the regulation came into force.

MirrorCaption's data architecture is relevant here:

• ✓
  No server-side audio storage — audio streams in real time through your browser to the transcription layer and is discarded after processing. MirrorCaption does not hold a copy of the interview audio on any server.
• ✓
  Local transcript storage — transcripts are saved in your browser's local storage (IndexedDB). They are not stored externally unless you choose to export them.
• ✓
  You control export and deletion — the transcript is your file. Export it, delete it, or keep it in the browser session. MirrorCaption does not retain it.

This removes the "vendor stores our interview audio" risk from your data landscape. The remaining compliance work falls on your side: candidate consent (required in most GDPR contexts), a defined retention policy for any transcripts you export, and access controls on who can view them. Those are standard HR data practices regardless of which tool you use.

Pricing — No Per-Seat Subscription

Most interview transcription tools charge per user per month. For teams running volume hiring, that adds up quickly. Unlike post-meeting tools such as Otter.ai, MirrorCaption's Premium tier is a one-time purchase, not a recurring subscription.

*200h hosted transcription credit included; additional hours via Voice Packs sold separately (€0.53–0.60/hr). All future product updates included. Competitor pricing and call-presence notes are summarized from vendor pages at time of publication; see each vendor's current documentation for details.

For a recruiter running 15 one-hour interviews a month, 200h of included credit covers more than a year of use before a top-up is needed. When the included credit runs out, Voice Packs add hours without a subscription. Premium customers get the lowest per-hour Voice Pack rate.

If you are not ready to commit, the free tier gives you 1 hour to run a real interview. No credit card, no monthly reset.

Frequently Asked Questions

Does the candidate need to install anything to use MirrorCaption in an interview?

No. MirrorCaption runs in the interviewer's browser only. The candidate joins the call as normal — they do not install or open anything extra. No browser extension is required on their end, and nothing changes about their meeting experience.

Can I use MirrorCaption for phone or in-person interviews?

Yes. Talk mode uses your device microphone, so it works for phone calls and face-to-face interviews. Open the app on your mobile device in Chrome, select Talk mode, and place the phone on the table or facing the speaker. The transcript and translation stream live on your screen throughout the conversation.

Which languages does MirrorCaption support for interview transcription?

MirrorCaption supports 50+ selectable languages including Mandarin, Japanese, Korean, Arabic, Hindi, Spanish, French, German, Portuguese, Russian, Hebrew, Turkish, Polish, and more. The source language (what the candidate speaks) and target language (what you read in the transcript) are selected independently, so an interviewer who speaks English can interview a candidate in Japanese and read a live English translation, while the candidate reads the interviewer's words in Japanese.

Does MirrorCaption record or store the interview audio?

No. Audio streams from your browser to the real-time transcription layer and is discarded after processing. MirrorCaption does not store interview audio on any server. Transcripts are saved locally in your browser session and are only exported if you choose to.

Is it GDPR-compliant to transcribe candidate interviews with MirrorCaption?

MirrorCaption does not store interview audio on external servers, which addresses a key GDPR data-minimisation concern. That said, you remain responsible for informing candidates that transcription is running — standard practice under GDPR regardless of the tool used. For current GDPR guidance on recording and transcribing job interviews, see the GDPR requirements for video interviews guide.

Start with One Interview

The interview is where you form most of the opinion that drives the hiring decision. Post-meeting tools are built for documentation. They are not built for the 45 minutes where the decision actually happens.

MirrorCaption is a browser tab that runs alongside your call. No install, no bot in the meeting, no upload wait. The candidate speaks. You read what they said in your language, while they are still talking. Speaker labels keep the transcript organized. When the interview ends, export the transcript and close the tab.

For international hiring teams conducting 10, 20, or 40 interviews a month across languages, that is the workflow. For teams that also manage ongoing multilingual collaboration beyond recruiting, see how live translation for sales calls applies the same approach to cross-border client conversations.