Smart contracts are increasingly used with blockchain systems for high-value applications. Blockchain’s immutability makes changing a deployed smart contract almost impossible, and any bug or security vulnerability in a deployed smart contract can have devastating consequences for developers and users of the smart contract. Thus, it is highly desired to ensure the quality of smart contracts before they are deployed, especially to detect all potential vulnerabilities in the smart contract source code. This paper proposes a new deep learning-based tool, MANDO-GURU, that aims to accurately detect vulnerabilities in smart contract source code at both coarse-grained contract-level and fine-grained line-level. Using a combination of control-flow graphs and call graphs of Solidity smart contracts, we design new heterogeneous graph attention neural networks to encode more structural and potentially semantic relations among different types of nodes and edges of such graphs and use the encoded embeddings of the graphs and nodes to detect vulnerabilities more accurately in Solidity code. Our validation of real-world smart contract datasets shows that MANDO-GURU can significantly improve many other vulnerability detection techniques by up to 24% in terms of F1-score at the contract level, depending on vulnerability types. More importantly, it is the first learning-based tool that identifies vulnerabilities at the line level and significantly improves the traditional code analysis-based techniques by up to 63.4%. Our tool is publicly available at https://github.com/MANDO-Project/ge-sc-machine. We provide a tutorial to use Docker Image to deploy on local machines in the Github repository. A test version is currently deployed at http://mandoguru.com, and a demo video of our tool is available at http://mandoguru.com/demo-video.