More than 350 hours of transcribed recordings—collected in various cities, across different communicative contexts, and involving people from diverse social backgrounds—totaling 994 speakers and approximately 3,200,000 tokens.
The corpus is modular and incremental: it was designed to grow over time, expanding the number of data collection sites, types of communicative situations, and speakers.
The transcripts are aligned with the audio and are freely accessible and searchable via the No SketchEngine search interface.
