A unified framework, VimTS, that leverages synergy between various text spotting tasks and scenarios to enhance the generalization ability of the model.