Transcribe audio

1. Projects → Create project. Name it on the Project Name tab.
2. On Labeling Setup, open Audio/Speech Processing and pick:
   • Automatic Speech Recognition — one text box per clip (the simplest setup);
   • Speech Transcription — mark Speech/Noise segments on the waveform and transcribe each segment; or
   • Automatic Speech Recognition using Segments — segment-based ASR correction.
3. On Data Import, provide tasks with an audio key pointing at your files. Drag-drop works for a proof of concept; for real datasets connect cloud storage so the originals stay in your bucket. Save.
4. Annotators play the clip (zoomable waveform, hotkey submit) and type the transcript. Tip: because the config has a TextArea, the Settings → Translation tab is available — its glossary can enforce consistent spellings of names and code-switched terms.

Go deeper

• Cloud storage → — the right way to import audio at scale
• Labeling guide → — the annotator's view
• Task lifecycle → — from imported to completed
• Export annotations → — transcripts out, incl. ASR-friendly formats

Add the languages present (code-mixed audio)

For code-mixed recordings, ask for every language heard without listing them by hand. The <Select> tag offers the complete ISO 639-3 registry, narrowed to a country or region, with search-as-you-type:

<Header value="Language(s) present"/>
<Select name="languages" toName="audio" dataset="languages" countries="NG"
        selected="English" codes="true" required="true"
        requiredMessage="Select every language you hear."/>

Add allowCustom="true" when annotators may meet a variety that is not in the registry, and extra="Naija" to pin a project-specific name.