The article presents a novel global fit pipeline for analyzing gravitational wave (GW) signals from Galactic binaries (GBs) and massive black hole binaries (MBHBs) in the LISA data. The pipeline is designed to robustly extract and characterize these signals while estimating the noise of the residual. The authors perform weekly analyses starting from one week to a full year of observation, demonstrating that the number of detected signals increases with longer observation times, bringing the noise estimate closer to the instrument noise level. They also introduce a maximum likelihood estimate (MLE) algorithm for extracting multiple MBHBs and discuss the impact of data gaps on MBHB signal extraction. The pipeline is evaluated using the LISA Data Challenge 2a (LDC2a) data set, which includes a combination of 15 MBHBs, 30 million GBs, and instrument noise. The results show a steady increase in the number of recovered signals and a stable match ratio of 85% to 88% after 16 weeks of observation. The pipeline is computationally efficient, with a cost of only 20 USD for a one-year analysis, and is publicly available for further research. Future work will focus on incorporating other GW signal types and extending the pipeline to include pre-merger MBHB parameter estimation.The article presents a novel global fit pipeline for analyzing gravitational wave (GW) signals from Galactic binaries (GBs) and massive black hole binaries (MBHBs) in the LISA data. The pipeline is designed to robustly extract and characterize these signals while estimating the noise of the residual. The authors perform weekly analyses starting from one week to a full year of observation, demonstrating that the number of detected signals increases with longer observation times, bringing the noise estimate closer to the instrument noise level. They also introduce a maximum likelihood estimate (MLE) algorithm for extracting multiple MBHBs and discuss the impact of data gaps on MBHB signal extraction. The pipeline is evaluated using the LISA Data Challenge 2a (LDC2a) data set, which includes a combination of 15 MBHBs, 30 million GBs, and instrument noise. The results show a steady increase in the number of recovered signals and a stable match ratio of 85% to 88% after 16 weeks of observation. The pipeline is computationally efficient, with a cost of only 20 USD for a one-year analysis, and is publicly available for further research. Future work will focus on incorporating other GW signal types and extending the pipeline to include pre-merger MBHB parameter estimation.