Report · October 2026 · second edition

What Latin America Transcribes

Two months of real JustTranscribe usage, in numbers: 6,683 recordings and videos turned into text by people in Argentina, Brazil, Peru, Colombia, Mexico and Chile. What they upload, how long it runs, in which language and at what hour.

UGBy Unai Goikoetxea·Data from 4 August to 5 October 2026 · second edition, published 6 October 2026 (first: 15 September)

More than half are files, not links

54% of transcripts start from a file someone had on their phone or computer. YouTube links are 35% of transcripts but almost half of all minutes: they are long lectures and talks. TikTok and Instagram add up to 8%, twice September's share, in one- or two-minute clips that barely register in total time.

More than a third of uploads are now video files

Of every ten uploaded files, almost four are video files (MP4/MOV), fewer than three are WhatsApp voice notes exactly as they leave the phone (.opus) and the rest are audio recordings: MP3s, phone memos, interviews, classes. In September the WhatsApp note was the most common format; in October it is video. Nobody converts formats before uploading; they upload what they have.

Half run under 7 minutes; the longest 19% take two thirds of the time

The typical recording is 6.7 minutes. But files over 30 minutes, fewer than one in five, hold 65% of all minutes transcribed: university lectures, meetings and hearings. 478 recordings ran past an hour, 194 past 90 minutes, and the longest hit the 150-minute ceiling.

Median length by source: Google Drive 38.7 min · YouTube 13.7 · uploaded file 3.9 · TikTok 1.5 · Instagram 1.2.

Spanish in almost two of every three; 57 languages in total

Language is detected automatically, never chosen by the user. Almost two in three recordings are Spanish, one in six English and one in nine Portuguese, which grew with Brazil's arrival. Then French, Arabic, Hindi, Chinese, Hebrew, Russian and Japanese: 57 distinct languages in two months.

The model labelled 56 recordings as Galician; on review most were Argentine or Chilean Spanish with a strong accent, a known error we fixed on 12 and 14 September. That is why Galician is in neither the list nor the language count.

Argentina first, Brazil now second, and Lima is still the city that transcribes most

Almost half of all transcripts start in Argentina. Brazil, which arrived in September with the Portuguese version, went from fourth to second country in three weeks, ahead of Peru and Colombia. By city, Lima leads, then Buenos Aires, Bogotá, Córdoba and Rosario; São Paulo is the first Brazilian city. Seven in ten transcripts are started from a computer.

Afternoons, and Mondays

The peak hour is 17:00 to 18:00 in Bogotá and Lima (19:00 to 20:00 in Buenos Aires), with a second peak around noon. Almost nobody transcribes between 3 and 5 in the morning. Monday is the busiest day, 45% above Saturday: what was recorded during the week gets turned into text as the next one starts.

Word wins: six of every ten exports

Six in ten accounts came back to transcribe at least once more and 11% made five or more transcripts; the most active account is at 75. When people export, it is Word 60% of the time (since 12 September it has its own button next to Copy); TXT, PDF and SRT subtitles share the rest almost equally, with subtitles rising. 60 transcripts were translated into another language, 17 languages in all. Speaker labels became automatic on 20 September, so their usage no longer compares with the first edition.

Export formats

What the file names say

We do not read content for this report; we do count which words appear in file names. WhatsApp appears in 1,040 names, class in 151, interview in 122 and meeting in 53. There are also podcasts and church sermons, few but steady.