Phonetic alignment is the task of finding the limits of phones and higher units in an audio file. This has been reliably done in many languages such as English, French and German, but, so far, no available Brazilian Portuguese aligner had a performance comparable with the ones used for these other languages. Thus, the main goal of this work was to implement a useful tool for forced alignment for Brazilian Portuguese. The implementation was done in two steps, the grapheme-to-phoneme conversion and the alignment itself. The Converter is responsible for receiving the input transcription in graphemes and converting it to its equivalent in phonemes and allophones, and was implemented using computational rules derived from the analysis of regular grapheme-phoneme relations in Brazilian Portuguese and an exception dictionary, for words to which no regular rules could be applied. The Aligner was responsible for aligning the phonemes/allophones of the previous module to the corresponding acoustic intervals of the audio file, called "phones". This module was implemented using hidden Markov models. Results for the Converter have an accuracy of over 99%, where the main mistakes involved mid vowels /e/ and /ɛ/ and /o/ and /ɔ/. As for the Aligner, the best model has 87% of the alignments with errors below 25 ms.