SpeechIndexer is a software that is used to segment and transcribe recorded speech. In particular, it allows to correlate sections of the audio with the corresponding textual transcription. Both the transcribing and indexing process is supported by an efficient pause finder that detects intonational units. The original purpose of the software was the documentation of endangered Formosan aboriginal languages. These languages have a pure oral tradition and it is only recently that archives of speech recordings in these languages exist. SpeechIndexer can be used also to teach and learn languages. Teaching material can be created from read and spontaneous speech, podcasts, etc.