diff --git a/NAMESPACE b/NAMESPACE
index 7171146..7a39c4a 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -12,3 +12,4 @@ export(dft_rast_transition)
 export(dft_stac_classes)
 export(dft_stac_config)
 export(dft_stac_fetch)
+export(dft_transition_vectors)
diff --git a/R/dft_transition_vectors.R b/R/dft_transition_vectors.R
new file mode 100644
index 0000000..108a7c7
--- /dev/null
+++ b/R/dft_transition_vectors.R
@@ -0,0 +1,144 @@
+#' Vectorize transition raster into individual change patches
+#'
+#' Convert a transition `SpatRaster` (from [dft_rast_transition()]) into `sf`
+#' polygons — one row per connected patch of pixels sharing the same
+#' transition type. Useful for QA in GIS, spatial attribution to management
+#' zones, and patch-level reporting.
+#'
+#' @param x A factor `SpatRaster` from [dft_rast_transition()] (the `$raster`
+#'   element). Must have a projected CRS.
+#' @param zones Optional `sf` polygon layer for spatial attribution. Any
+#'   partitioning: sub-basins, parcels, climate regions, management units.
+#' @param zone_col Character. Column name in `zones` identifying each zone.
+#'   Required when `zones` is supplied.
+#' @param patch_area_min Numeric or `NULL`. Minimum patch area in m². Patches
+#'   smaller than this are dropped before returning. `NULL` (default) keeps all.
+#'
+#' @return An `sf` data frame (polygon geometry) with columns:
+#'   - `patch_id` (integer) — connected component ID
+#'   - `transition` (character) — transition label (e.g. "Trees -> Rangeland")
+#'   - `area_ha` (numeric) — patch area in hectares
+#'   - Zone column (if `zones` supplied) — from spatial intersection
+#'
+#' @export
+#' @examples
+#' r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+#' r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+#' classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+#' result <- dft_rast_transition(classified, from = "2017", to = "2020")
+#'
+#' # Vectorize all transition patches
+#' patches <- dft_transition_vectors(result$raster)
+#' head(patches)
+#'
+#' # Filter to large patches only
+#' patches_large <- dft_transition_vectors(result$raster, patch_area_min = 1000)
+#' head(patches_large)
+dft_transition_vectors <- function(x,
+                                   zones = NULL,
+                                   zone_col = NULL,
+                                   patch_area_min = NULL) {
+  if (!inherits(x, "SpatRaster")) {
+    stop("`x` must be a SpatRaster (e.g. from dft_rast_transition()$raster).",
+         call. = FALSE)
+  }
+  dft_check_crs(x, "dft_transition_vectors")
+
+  if (!terra::is.factor(x)) {
+    stop("`x` must be a factor SpatRaster with transition labels.", call. = FALSE)
+  }
+
+  if (!is.null(zones)) {
+    if (!inherits(zones, "sf")) {
+      stop("`zones` must be an sf object.", call. = FALSE)
+    }
+    if (is.null(zone_col) || !zone_col %in% names(zones)) {
+      stop("`zone_col` must name a column in `zones`.", call. = FALSE)
+    }
+  }
+
+  if (!is.null(patch_area_min)) {
+    if (!is.numeric(patch_area_min) || length(patch_area_min) != 1 ||
+          is.na(patch_area_min) || patch_area_min < 0) {
+      stop("`patch_area_min` must be a single non-negative number or NULL.",
+           call. = FALSE)
+    }
+  }
+
+  lvls <- terra::cats(x)[[1]]
+  vals <- terra::values(x)[, 1]
+
+  if (all(is.na(vals))) {
+    return(sf::st_sf(
+      patch_id = integer(0), transition = character(0),
+      area_ha = numeric(0), geometry = sf::st_sfc(crs = sf::st_crs(x))
+    ))
+  }
+
+  # Build unique patch IDs per transition type
+  patch_ids <- rep(NA_integer_, terra::ncell(x))
+  patch_transition <- integer(0)
+  offset <- 0L
+
+  for (i in seq_len(nrow(lvls))) {
+    code <- lvls$id[i]
+    mask <- which(vals == code)
+    if (length(mask) == 0) next
+
+    r_mask <- terra::rast(x)
+    mask_vals <- rep(NA_integer_, terra::ncell(x))
+    mask_vals[mask] <- 1L
+    terra::values(r_mask) <- mask_vals
+    p <- terra::patches(r_mask, directions = 8)
+    p_vals <- terra::values(p)[, 1]
+
+    valid <- !is.na(p_vals)
+    unique_pids <- sort(unique(p_vals[valid]))
+    for (pid in unique_pids) {
+      offset <- offset + 1L
+      patch_ids[valid & p_vals == pid] <- offset
+      patch_transition <- c(patch_transition, code)
+    }
+  }
+
+  if (offset == 0L) {
+    return(sf::st_sf(
+      patch_id = integer(0), transition = character(0),
+      area_ha = numeric(0), geometry = sf::st_sfc(crs = sf::st_crs(x))
+    ))
+  }
+
+  # Vectorize patch IDs
+  r_pid <- terra::rast(x)
+  names(r_pid) <- "pid"
+  terra::values(r_pid) <- patch_ids
+  polys <- terra::as.polygons(r_pid)
+  polys_sf <- sf::st_as_sf(polys)
+
+  # Map patch IDs to transition labels
+  code_to_label <- stats::setNames(lvls$transition, lvls$id)
+  polys_sf$patch_id <- as.integer(polys_sf$pid)
+  polys_sf$transition <- as.character(
+    code_to_label[as.character(patch_transition[polys_sf$pid])]
+  )
+  polys_sf$area_ha <- as.numeric(sf::st_area(polys_sf)) * 1e-4
+  polys_sf$pid <- NULL
+
+  # Filter by minimum area
+  if (!is.null(patch_area_min) && patch_area_min > 0) {
+    area_m2 <- as.numeric(sf::st_area(polys_sf))
+    polys_sf <- polys_sf[area_m2 >= patch_area_min, ]
+  }
+
+  # Select and order columns
+  out <- polys_sf[c("patch_id", "transition", "area_ha")]
+
+  # Zone attribution
+
+  if (!is.null(zones)) {
+    zones_proj <- sf::st_transform(zones[zone_col], sf::st_crs(out))
+    out <- suppressWarnings(sf::st_intersection(out, zones_proj))
+  }
+
+  out
+}
diff --git a/man/dft_transition_vectors.Rd b/man/dft_transition_vectors.Rd
new file mode 100644
index 0000000..95d568b
--- /dev/null
+++ b/man/dft_transition_vectors.Rd
@@ -0,0 +1,50 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/dft_transition_vectors.R
+\name{dft_transition_vectors}
+\alias{dft_transition_vectors}
+\title{Vectorize transition raster into individual change patches}
+\usage{
+dft_transition_vectors(x, zones = NULL, zone_col = NULL, patch_area_min = NULL)
+}
+\arguments{
+\item{x}{A factor \code{SpatRaster} from \code{\link[=dft_rast_transition]{dft_rast_transition()}} (the \verb{$raster}
+element). Must have a projected CRS.}
+
+\item{zones}{Optional \code{sf} polygon layer for spatial attribution. Any
+partitioning: sub-basins, parcels, climate regions, management units.}
+
+\item{zone_col}{Character. Column name in \code{zones} identifying each zone.
+Required when \code{zones} is supplied.}
+
+\item{patch_area_min}{Numeric or \code{NULL}. Minimum patch area in m². Patches
+smaller than this are dropped before returning. \code{NULL} (default) keeps all.}
+}
+\value{
+An \code{sf} data frame (polygon geometry) with columns:
+\itemize{
+\item \code{patch_id} (integer) — connected component ID
+\item \code{transition} (character) — transition label (e.g. "Trees -> Rangeland")
+\item \code{area_ha} (numeric) — patch area in hectares
+\item Zone column (if \code{zones} supplied) — from spatial intersection
+}
+}
+\description{
+Convert a transition \code{SpatRaster} (from \code{\link[=dft_rast_transition]{dft_rast_transition()}}) into \code{sf}
+polygons — one row per connected patch of pixels sharing the same
+transition type. Useful for QA in GIS, spatial attribution to management
+zones, and patch-level reporting.
+}
+\examples{
+r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+result <- dft_rast_transition(classified, from = "2017", to = "2020")
+
+# Vectorize all transition patches
+patches <- dft_transition_vectors(result$raster)
+head(patches)
+
+# Filter to large patches only
+patches_large <- dft_transition_vectors(result$raster, patch_area_min = 1000)
+head(patches_large)
+}
diff --git a/tests/testthat/test-dft_transition_vectors.R b/tests/testthat/test-dft_transition_vectors.R
new file mode 100644
index 0000000..6bcde52
--- /dev/null
+++ b/tests/testthat/test-dft_transition_vectors.R
@@ -0,0 +1,109 @@
+test_that("dft_transition_vectors returns sf with expected columns", {
+  r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+  classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+  result <- dft_rast_transition(classified, from = "2017", to = "2020")
+
+  patches <- dft_transition_vectors(result$raster)
+
+  expect_s3_class(patches, "sf")
+  expect_true(all(c("patch_id", "transition", "area_ha") %in% names(patches)))
+  expect_true(nrow(patches) > 0)
+})
+
+test_that("total area matches raster summary", {
+  r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+  classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+  result <- dft_rast_transition(classified, from = "2017", to = "2020")
+
+  patches <- dft_transition_vectors(result$raster)
+
+  expect_equal(sum(patches$area_ha), sum(result$summary$area), tolerance = 0.1)
+})
+
+test_that("each patch has a valid transition label", {
+  r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+  classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+  result <- dft_rast_transition(classified, from = "2017", to = "2020")
+
+  patches <- dft_transition_vectors(result$raster)
+  lvls <- terra::cats(result$raster)[[1]]
+
+  expect_true(all(patches$transition %in% lvls$transition))
+})
+
+test_that("patch_ids are unique", {
+  r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+  classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+  result <- dft_rast_transition(classified, from = "2017", to = "2020")
+
+  patches <- dft_transition_vectors(result$raster)
+
+  expect_equal(length(unique(patches$patch_id)), nrow(patches))
+})
+
+test_that("patch_area_min filters small patches", {
+  r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+  classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+  result <- dft_rast_transition(classified, from = "2017", to = "2020")
+
+  all_patches <- dft_transition_vectors(result$raster)
+  filtered <- dft_transition_vectors(result$raster, patch_area_min = 1000)
+
+  expect_lt(nrow(filtered), nrow(all_patches))
+  expect_true(all(as.numeric(sf::st_area(filtered)) >= 1000))
+})
+
+test_that("zone attribution adds zone column", {
+  r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+  aoi <- sf::st_read(
+    system.file("extdata", "example_aoi.gpkg", package = "drift"), quiet = TRUE
+  )
+  classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+  result <- dft_rast_transition(classified, from = "2017", to = "2020")
+
+  # Use AOI as a single zone
+  aoi$zone_name <- "test_zone"
+  patches <- dft_transition_vectors(result$raster, zones = aoi,
+                                    zone_col = "zone_name")
+
+  expect_true("zone_name" %in% names(patches))
+})
+
+test_that("errors on non-SpatRaster input", {
+  expect_error(dft_transition_vectors(data.frame(x = 1)),
+               "SpatRaster")
+})
+
+test_that("errors on non-factor SpatRaster", {
+  r <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  expect_error(dft_transition_vectors(r), "factor")
+})
+
+test_that("errors on geographic CRS", {
+  r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+  classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+  result <- dft_rast_transition(classified, from = "2017", to = "2020")
+  r_geo <- terra::project(result$raster, "EPSG:4326", method = "near")
+
+  expect_error(dft_transition_vectors(r_geo), "projected CRS")
+})
+
+test_that("errors when zones supplied without zone_col", {
+  r17 <- terra::rast(system.file("extdata", "example_2017.tif", package = "drift"))
+  r20 <- terra::rast(system.file("extdata", "example_2020.tif", package = "drift"))
+  classified <- dft_rast_classify(list("2017" = r17, "2020" = r20), source = "io-lulc")
+  result <- dft_rast_transition(classified, from = "2017", to = "2020")
+  aoi <- sf::st_read(
+    system.file("extdata", "example_aoi.gpkg", package = "drift"), quiet = TRUE
+  )
+
+  expect_error(dft_transition_vectors(result$raster, zones = aoi),
+               "zone_col")
+})