what MAXQDA expects
MAXQDA's transcript import has a particular format for synchronized audio: timestamps in the form #00:01:23.4# inserted inline at each speaker turn, with named speakers (not generic labels), and paragraph breaks at every turn so the auto-segmentation step has clean edges.
a generic .docx imports — and then you spend an hour adding timestamp markers, splitting paragraphs at speaker switches, and replacing "Speaker 1" with the actual participant code. for a 25-interview study, that's a billable week.
what we ship
- inline timestamp markers in MAXQDA's expected format:
#hh:mm:ss.s#at every speaker turn. matches MAXQDA's transcript-audio sync import pattern. - named speakers from your codebook. relabel once in the editor, propagate through every row.
- per-turn paragraph breaks so MAXQDA's auto-coding pass has clean turn boundaries to work with.
- .docx export with the correct styling so MAXQDA's import wizard parses it without manual cleanup. plain text and json exports for researchers who code in separate documents or build their own import pipelines.
workflow
- drop the interview audio. transcription runs with diarization.
- review in the editor. relabel speakers to your codebook. verify proper nouns where needed.
- export to MAXQDA format. open MAXQDA, create a project, import the transcript with the audio file linked. timestamps sync automatically.
- code as normal. clicking a coded segment plays the audio from that timestamp.
where this fits
- fits: mixed-methods studies, focus groups, ethnographic interviews. the use cases MAXQDA is built around — same use cases that work in NVivo and ATLAS.ti.
- fits: studies that already use MAXQDA's intercoder agreement features. clean per-turn structure makes the kappa computation cleaner. no spurious segment boundaries.
- doesn't fit (yet): MAXQDA's MAXMaps for visualizing relationships across coded segments. our export is the coding substrate; the visualization happens after coding, in MAXQDA itself.
privacy
for fieldwork audio that can't go through a cloud vendor — IRB-restricted, GDPR-bound, on-the-record-but-not-public — run the file in private mode. the MAXQDA export works identically; the audio just stays on your laptop.
citation
methodology footnote we recommend: "transcribed using audiohighlight v[X], MAXQDA export profile, on [date]." we keep version numbers stable for replication.