Background: Liver fibrosis is a common feature in chronic liver disease, and an underlying condition in hepatocellular carcinoma (HCC). Liver biopsy, the gold-standard to assess fibrosis, is not suitable for screening of at-risk populations. Thus, the use of non-invasive assessment tools such as ultrasound and biomarkers have gained attention. We investigated the utility of salivary biomarkers to detect cirrhosis. We developed the Saliva Liver Fibrosis (SALF) score, the first saliva-based algorithm for the screening and/or early diagnosis of liver fibrosis. Methods: Saliva samples from 135 patients were collected: liver cirrhosis (n=41), intermediate stages of liver fibrosis (n=20), liver disease with minimal fibrosis (n=50), and healthy controls (n=41). Liver fibrosis stage was determined by liver stiffness measurement (LSM) using transient elastography. Individuals with LSM<7.0 kPa were considered to have minimal/nil fibrosis, LSM of 8.0 kPa-12.0 kPa was categorized as intermediate fibrosis, and liver cirrhosis was classified as LSM>14.0 kPa. Clinically used serum biomarkers (hyaluronic acid, HA, tissue inhibitor of metalloproteinase 1, TIMP-1, and α-2-macroglobulin, A2MG) were measured in saliva samples, and then utilized in a logistic regression analysis. Results: Salivary concentrations of HA, TIMP-1 and A2MG were increased (p<0.05) in patients with liver fibrosis/cirrhosis compared to those without fibrosis. By combining these three biomarkers, we developed a diagnostic score to detect significant liver fibrosis, the SALF score. In a training cohort, the median SALF scores of patients with liver cirrhosis (0.921±0.09) and fibrosis (0.819±0.285) were significantly higher than the scores of patients with minimal/nil fibrosis (0.061±0.29) and healthy controls (0.034±0.05). The SALF score identified patients with significant liver fibrosis with an Area Under the Curve (AUC) of 0.970, with a performance that was similar or superior to clinically validated diagnostic algorithms such as the Fibrosis-4 (FIB-4, AUC: 0.740) and Hepascore (AUC: 0.979). Using the optimal cut-off of 0.51, the validation of the SALF score in an independent cohort showed an AUC of 0.920 for the detection of significant liver fibrosis. Conclusion: To our knowledge, this is the first study to demonstrate clinical utility of saliva to diagnose liver fibrosis/ cirrhosis. The SALF score is a novel and accurate tool to improve the screening for cirrhosis in asymptomatic populations.