diff --git a/src/iceberg/inspect/metadata_scan_util_internal.h b/src/iceberg/inspect/metadata_scan_util_internal.h
index 5cd2058f9..65e1f3ef3 100644
--- a/src/iceberg/inspect/metadata_scan_util_internal.h
+++ b/src/iceberg/inspect/metadata_scan_util_internal.h
@@ -40,6 +40,7 @@
 #include "iceberg/schema_internal.h"
 #include "iceberg/type.h"
 #include "iceberg/util/macros.h"
+#include "iceberg/util/stream.h"
 
 namespace iceberg::internal {
 
@@ -181,4 +182,79 @@ Result<ArrowArrayStream> MakeMetadataBatchStream(
                                batch_size, std::forward<GetRow>(get_row)));
 }
 
+/// Stream manifest-backed rows with the same batch-size and ownership contract.
+template <typename Row, typename AppendRow>
+class MetadataRowsSource : public Stream<ArrowArray> {
+ public:
+  MetadataRowsSource(std::shared_ptr<iceberg::Schema> schema,
+                     std::unique_ptr<Stream<Row>> rows, int64_t batch_size,
+                     AppendRow append_row)
+      : schema_(std::move(schema)),
+        rows_(std::move(rows)),
+        batch_size_(batch_size),
+        append_row_(std::move(append_row)) {}
+
+  Status Close() {
+    rows_.reset();
+    return {};
+  }
+
+  Result<ArrowSchema> Schema() {
+    ArrowSchema schema{};
+    ICEBERG_RETURN_UNEXPECTED(ToArrowSchema(*schema_, &schema));
+    return schema;
+  }
+
+ protected:
+  Result<std::optional<ArrowArray>> NextImpl() override {
+    ICEBERG_PRECHECK(rows_ != nullptr, "Cannot read from a closed metadata stream");
+    ICEBERG_ASSIGN_OR_RAISE(auto row, rows_->Next());
+    if (!row) {
+      return std::nullopt;
+    }
+    ICEBERG_ASSIGN_OR_RAISE(auto builder, ArrowRowBuilder::Make(*schema_));
+    ICEBERG_RETURN_UNEXPECTED(append_row_(builder, *row));
+    while (builder.num_rows() < batch_size_) {
+      ICEBERG_ASSIGN_OR_RAISE(auto next_row, rows_->Next());
+      if (!next_row) {
+        break;
+      }
+      ICEBERG_RETURN_UNEXPECTED(append_row_(builder, *next_row));
+    }
+    ICEBERG_ASSIGN_OR_RAISE(auto batch, std::move(builder).Finish());
+    return batch;
+  }
+
+ private:
+  const std::shared_ptr<iceberg::Schema> schema_;
+  std::unique_ptr<Stream<Row>> rows_;
+  const int64_t batch_size_;
+  AppendRow append_row_;
+};
+
+template <typename Row, typename AppendRow>
+Result<ArrowArrayStream> MakeMetadataBatchStream(const std::shared_ptr<Schema>& schema,
+                                                 std::unique_ptr<Stream<Row>> rows,
+                                                 int64_t batch_size,
+                                                 AppendRow append_row) {
+  ICEBERG_PRECHECK(batch_size > 0, "Batch size must be positive");
+  ICEBERG_PRECHECK(schema != nullptr, "Metadata stream schema cannot be null");
+  ICEBERG_PRECHECK(rows != nullptr, "Metadata row source cannot be null");
+  if constexpr (requires { static_cast<bool>(append_row); }) {
+    ICEBERG_PRECHECK(static_cast<bool>(append_row),
+                     "Metadata row callback cannot be empty");
+  }
+  return MakeArrowArrayStream(std::make_unique<MetadataRowsSource<Row, AppendRow>>(
+      schema, std::move(rows), batch_size, std::move(append_row)));
+}
+
+template <typename Row, typename AppendRow>
+Result<ArrowArrayStream> MakeMetadataBatchStream(const std::shared_ptr<Schema>& schema,
+                                                 std::vector<Row> rows,
+                                                 int64_t batch_size,
+                                                 AppendRow append_row) {
+  return MakeMetadataBatchStream(schema, Stream<Row>::FromVector(std::move(rows)),
+                                 batch_size, std::move(append_row));
+}
+
 }  // namespace iceberg::internal
diff --git a/src/iceberg/test/CMakeLists.txt b/src/iceberg/test/CMakeLists.txt
index 1874fc25e..ec4721635 100644
--- a/src/iceberg/test/CMakeLists.txt
+++ b/src/iceberg/test/CMakeLists.txt
@@ -206,6 +206,7 @@ if(ICEBERG_BUILD_BUNDLE)
   add_iceberg_test(metadata_table_test
                    USE_BUNDLE
                    SOURCES
+                   inspect/metadata_rows_test.cc
                    inspect/history_table_test.cc
                    inspect/metadata_table_test.cc
                    inspect/refs_table_test.cc
diff --git a/src/iceberg/test/inspect/metadata_rows_test.cc b/src/iceberg/test/inspect/metadata_rows_test.cc
new file mode 100644
index 000000000..be69880a0
--- /dev/null
+++ b/src/iceberg/test/inspect/metadata_rows_test.cc
@@ -0,0 +1,137 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements.  See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership.  The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License.  You may obtain a copy of the License at
+ *
+ *   http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing,
+ * software distributed under the License is distributed on an
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ * KIND, either express or implied.  See the License for the
+ * specific language governing permissions and limitations
+ * under the License.
+ */
+
+#include <cerrno>
+#include <cstdint>
+#include <functional>
+#include <memory>
+#include <optional>
+#include <utility>
+#include <vector>
+
+#include <arrow/array.h>
+#include <arrow/type.h>
+#include <gtest/gtest.h>
+#include <nanoarrow/nanoarrow.hpp>
+
+#include "iceberg/inspect/metadata_scan_util_internal.h"
+#include "iceberg/inspect/metadata_table.h"
+#include "iceberg/partition_spec.h"
+#include "iceberg/schema.h"
+#include "iceberg/schema_field.h"
+#include "iceberg/test/inspect/metadata_table_test_base.h"
+#include "iceberg/test/matchers.h"
+#include "iceberg/transform.h"
+#include "iceberg/type.h"
+
+namespace iceberg::internal {
+namespace {
+
+Status AppendIntRow(ArrowRowBuilder& builder, const int64_t& row) {
+  ICEBERG_RETURN_UNEXPECTED(AppendInt(builder.column(0), row));
+  return builder.FinishRow();
+}
+
+std::shared_ptr<Schema> IntSchema() {
+  return std::make_shared<Schema>(
+      std::vector<SchemaField>{SchemaField::MakeRequired(1, "id", int64())});
+}
+
+class CountingRows : public Stream<int64_t> {
+ public:
+  explicit CountingRows(std::shared_ptr<int> reads) : reads_(std::move(reads)) {}
+
+ protected:
+  Result<std::optional<int64_t>> NextImpl() override {
+    ++*reads_;
+    if (next_ == 5) {
+      return std::nullopt;
+    }
+    return next_++;
+  }
+
+ private:
+  std::shared_ptr<int> reads_;
+  int64_t next_ = 0;
+};
+
+TEST(MetadataRowsTest, RejectsInvalidInputsBeforeExport) {
+  EXPECT_THAT(MakeMetadataBatchStream(IntSchema(), std::unique_ptr<Stream<int64_t>>{}, 2,
+                                      AppendIntRow),
+              IsError(ErrorKind::kInvalidArgument));
+  EXPECT_THAT(MakeMetadataBatchStream(std::shared_ptr<Schema>{}, std::vector<int64_t>{1},
+                                      2, AppendIntRow),
+              IsError(ErrorKind::kInvalidArgument));
+  EXPECT_THAT(
+      MakeMetadataBatchStream(IntSchema(), std::vector<int64_t>{1}, 0, AppendIntRow),
+      IsError(ErrorKind::kInvalidArgument));
+  decltype(&AppendIntRow) callback = nullptr;
+  EXPECT_THAT(MakeMetadataBatchStream(IntSchema(), std::vector<int64_t>{1}, 2, callback),
+              IsError(ErrorKind::kInvalidArgument));
+  std::function<Status(ArrowRowBuilder&, const int64_t&)> empty;
+  EXPECT_THAT(MakeMetadataBatchStream(IntSchema(), std::vector<int64_t>{1}, 2, empty),
+              IsError(ErrorKind::kInvalidArgument));
+}
+
+TEST(MetadataRowsTest, LazyBatchesAndExhaustion) {
+  auto reads = std::make_shared<int>(0);
+  std::unique_ptr<Stream<int64_t>> rows = std::make_unique<CountingRows>(reads);
+  ICEBERG_UNWRAP_OR_FAIL(auto stream, MakeMetadataBatchStream(
+                                          IntSchema(), std::move(rows), 2, AppendIntRow));
+  nanoarrow::UniqueArrayStream owned(&stream);
+  nanoarrow::UniqueArray batch;
+  EXPECT_EQ(*reads, 0);
+  for (int length : {2, 2, 1}) {
+    ASSERT_EQ(owned->get_next(owned.get(), batch.get()), 0);
+    ASSERT_NE(batch->release, nullptr);
+    EXPECT_EQ(batch->length, length);
+    batch.reset();
+  }
+  for (int i = 0; i < 3; ++i) {
+    EXPECT_EQ(owned->get_next(owned.get(), batch.get()), 0);
+    EXPECT_EQ(batch->release, nullptr);
+  }
+  EXPECT_EQ(*reads, 6);
+}
+
+TEST(MetadataRowsTest, AppendFailureIsTerminal) {
+  auto reads = std::make_shared<int>(0);
+  std::unique_ptr<Stream<int64_t>> rows = std::make_unique<CountingRows>(reads);
+  auto append = [](ArrowRowBuilder& builder, const int64_t& row) -> Status {
+    if (row == 3) {
+      return InvalidArrowData("Row append failed");
+    }
+    return AppendIntRow(builder, row);
+  };
+  ICEBERG_UNWRAP_OR_FAIL(
+      auto stream, MakeMetadataBatchStream(IntSchema(), std::move(rows), 2, append));
+  nanoarrow::UniqueArrayStream owned(&stream);
+  nanoarrow::UniqueArray batch;
+  ASSERT_EQ(owned->get_next(owned.get(), batch.get()), 0);
+  batch.reset();
+  for (int i = 0; i < 3; ++i) {
+    EXPECT_EQ(owned->get_next(owned.get(), batch.get()), EIO);
+    EXPECT_EQ(batch->release, nullptr);
+    EXPECT_STREQ(owned->get_last_error(owned.get()), "Row append failed");
+  }
+  EXPECT_EQ(*reads, 4);
+}
+
+}  // namespace
+}  // namespace iceberg::internal