Estimating CATE using random causal forest with only means and variances of different features

I am planing to estimate CATE for different segments in my dataset using random causal forest. The data on the individual level is not available but the means and variances for each feature as well as features combination are available. My understanding from the original paper is that the split at each level of the tree is based on an objective function which only requires of the means, variances and number of individuals for each features for the test and control groups and thus I wonder if I can access an API for this particular case. I appreciate any help on this.

Start by comparing the original paper's random causal forest split requirements with EconML's available API and the stated means, variances, feature combinations, and group sizes. Determine whether an existing entry point supports aggregate data; done would be a clear API path or a documented conclusion that individual-level data is required.