The cellRanger output should also contain the matrix, which can be read in with Seurat's read10X() function

It says in the code, object.R:

Convert.seurat <- function(
  from,
  to,
  filename,
  chunk.dims = 'auto',
  chunk.size = 1000,
  overwrite = FALSE,
  display.progress = TRUE,
  anndata.raw = "raw.data",
  anndata.X = "data",
  ...
) {
  object.to <- switch(
    EXPR = to,
    'loom' = {
      stop("Converting to and from loom files is currently unavailable; we are working on restoring this functionality")
      # if (!PackageCheck('loomR', error = FALSE)) {
      #   stop("Please install loomR from GitHub before converting to a loom object")
      # }
      # # cell.order <- from@cell.names
      # cell.order <- colnames(x = from)
      # gene.order <- rownames(x = from)
      # loomfile <- loomR::create(
      #   filename = filename,
      #   data = GetAssayData(object = from, slot = 'counts')[, cell.order],
      #   cell.attrs = from[[]][cell.order, ],
      #   layers = list('norm_data' = t(x = GetAssayData(object = from)[, cell.order])),
      #   chunk.dims = chunk.dims,
      #   chunk.size = chunk.size,
      #   overwrite = overwrite,
      #   display.progress = display.progress
      # )
      # if (nrow(x = HVFInfo(object = from)) > 0) {
      #   hvg.info <- HVFInfo(object = from)
      #   colnames(x = hvg.info) <- gsub(
      #     pattern = '.',
      #     replacement = '_',
      #     x = colnames(x = hvg.info),
      #     fixed = TRUE
      #   )
...