Audio-to-lyrics transcription and alignment requires strong acoustic and language models that are trained on in-domain data and a well-adapted pronunciation model for singing. Even in the presence of such models, the length of audio segments for decoding remains a challenge. In this year’s MIREX submission, we present a recursive search method that splits the audio with respect to anchor- ing words for performing alignment on shorter audio seg- ments. The recursive is applied through gradually restrict- ing the language model and search space after each search iteration. We apply a final pass of forced alignment on the segmented audio to obtain timings for every word in the input song lyrics. According to the initial experiments, our system is robust to various musical genre while being executable on local machines with low memory and com- putational resources.