Turn survey audio into text that stays attached to the right answer.

Your survey has recorded open-ended answers, but a folder of text files is not enough for coding or analysis. Each response needs the correct project, respondent and question. Define that linkage at the start so your team can reconcile the export with the survey records and inspect mapping exceptions.

Scope an export where every transcript row carries its supplied keys and source filename. Preserve those identifiers when a translation is added, so the source-language text and translated response can remain paired. If a question key is missing or ambiguous, flag the mapping for review instead of guessing where the answer belongs.

Verbatim text preserves spoken wording, including repetitions where required. Cleaned text improves readability under agreed editing rules. Translation creates text in the requested output language. These choices should preserve meaning and remain explicit; a short summary should not silently replace the respondent’s full answer.

A question-specific CAPI recording may map directly to one response. A longer CATI interview may need question segmentation, speaker handling and consideration of answers given out of order. Evaluate those input formats separately, using the actual recording structure rather than assuming that one mapping approach fits every survey.

Agree how to report missing files, duplicate filenames, unclear speech and language changes. Define which passages need correction, who reviews them and how corrected text returns to the final export. Reconcile input records and output rows so your team can see what was processed and what still needs attention.