Background: Early cancer detection is key to reducing cancer deaths. Unfortunately, many established cancer screening technologies are not suitable for use in low- and middle-income countries due to expenses, complexity, and dependency on extensive medical infrastructure. Therefore, a simple, affordable, efficient multi-cancer early detection test is a major unmet need. Methods: This study enrolled nearly 10,000 participants (1959 cancer cases and 7423 non-cancer cases) containing more than nine common cancer types. One tube of peripheral blood samples was collected from each participant and quantified a panel of seven selected protein tumor markers (PTMs) by a clinical common electrochemiluminescence immunoassay analyzer. A protein assay named OncoSeek was established using artificial intelligence to distinguish cancer from non-cancer cases by calculating the probability of cancer (POC) index based on the quantification results of 7 PTMs and clinical information including gender and age of the subjects. Then using another model to predict the possible affected tissue of origin (TOO) who has been detected with cancer signal. Results: In this study, we found that the conventional clinical method relied only on a single threshold for each protein tumor marker which would make a big problem that was when combining the result of those markers, the false positive rate would accumulate as the number of tests increased. Nevertheless, OncoSeek was empowered by artificial intelligence technology to significantly reduce the false positive rate, increasing the specificity from 56.9% (95% confidence interval (CI): 55.8% to 58.0%) to 92.9% (95% CI: 92.3% to 93.5%). In all cancer types sensitivity of the OncoSeek test was 51.7% (95% CI: 49.4% to 53.9%). The sensitivities ranged from 37.1% to 77.6% for the detection of the nine common cancer types (breast, colorectum, esophagus, liver, lung, lymphoma, ovary, pancreas, stomach), which account for ~59.2% of global cancer deaths annually. Furthermore, it has shown excellent sensitivity in several high-mortality cancer types for which there are a lack of routine screening tests in the clinic, such as the sensitivity of pancreatic cancer was 77.6% (95% CI: 69.3% to 84.6%). The overall accuracy of tissue of origin (TOO) prediction in true positives was 66.8%, which reduced the clinical diagnostic workup. Conclusions: In summary, this study supported that OncoSeek significantly outperforms the conventional clinical method, representing a novel blood-based test for multi-cancer early detection which is a non-invasive, easier, and efficient approach. Moreover, the accuracy of TOO of it which facilitates the follow-up diagnostic workup. As well as this method is affordable (the cost of the test is less than $25) and requires nothing more than a blood draw at the screening sites, which makes it more practical in low- and middle-income countries.