The authors introduce CAR-Toner, an AI-driven tool for predicting and optimizing CAR tonic signaling. Tonic signaling is crucial for CAR-T cell function, with inefficient signaling leading to poor persistence and excessive signaling causing exhaustion. Previous methods for quantifying positively charged patches (PCPs) on CAR antigen-binding domains were limited by reliance on external servers, slow processing, and lack of batch capabilities. CAR-Toner addresses these limitations by integrating protein databases, structural biology, and deep learning models. It uses a large dataset of 170,000 protein sequences and the ESM2 model for PCP prediction and optimization. CAR-Toner accurately predicts PCP scores and provides optimization strategies for improving CAR-T cell function. Testing showed strong correlation between AI-predicted and traditional PCP scores. The tool was used to optimize a CLL1 antibody, reducing its PCP score to improve CAR-T cell expansion and reduce exhaustion. The study also compared PCP features of different CAR antigen-binding domains (scFv, VHH, VNAR, VLR), finding that VHH-based CARs have the highest proportion within the optimal PCP range. CAR-Toner offers a powerful tool for optimizing CAR design, with potential to advance CAR-T therapy. The authors note that further validation is needed to assess the impact of PCP optimization on antigen-binding specificity and affinity. The study highlights the importance of PCP in CAR-T cell function and suggests that VHH-based CARs may be more effective. The work was supported by multiple funding sources and includes contributions from multiple institutions. The authors declare no competing interests.The authors introduce CAR-Toner, an AI-driven tool for predicting and optimizing CAR tonic signaling. Tonic signaling is crucial for CAR-T cell function, with inefficient signaling leading to poor persistence and excessive signaling causing exhaustion. Previous methods for quantifying positively charged patches (PCPs) on CAR antigen-binding domains were limited by reliance on external servers, slow processing, and lack of batch capabilities. CAR-Toner addresses these limitations by integrating protein databases, structural biology, and deep learning models. It uses a large dataset of 170,000 protein sequences and the ESM2 model for PCP prediction and optimization. CAR-Toner accurately predicts PCP scores and provides optimization strategies for improving CAR-T cell function. Testing showed strong correlation between AI-predicted and traditional PCP scores. The tool was used to optimize a CLL1 antibody, reducing its PCP score to improve CAR-T cell expansion and reduce exhaustion. The study also compared PCP features of different CAR antigen-binding domains (scFv, VHH, VNAR, VLR), finding that VHH-based CARs have the highest proportion within the optimal PCP range. CAR-Toner offers a powerful tool for optimizing CAR design, with potential to advance CAR-T therapy. The authors note that further validation is needed to assess the impact of PCP optimization on antigen-binding specificity and affinity. The study highlights the importance of PCP in CAR-T cell function and suggests that VHH-based CARs may be more effective. The work was supported by multiple funding sources and includes contributions from multiple institutions. The authors declare no competing interests.