diff --git a/lib/CMakeLists.txt b/lib/CMakeLists.txt
index b3588fbf6..94e6b7b35 100644
--- a/lib/CMakeLists.txt
+++ b/lib/CMakeLists.txt
@@ -361,6 +361,7 @@ if(NOT EMSCRIPTEN)
       ${CMAKE_SOURCE_DIR}/test/json_dataview_test.cc
       ${CMAKE_SOURCE_DIR}/test/memory_filesystem_test.cc
       ${CMAKE_SOURCE_DIR}/test/parquet_test.cc
+      ${CMAKE_SOURCE_DIR}/test/prepared_statement_test.cc
       ${CMAKE_SOURCE_DIR}/test/readahead_buffer_test.cc
       ${CMAKE_SOURCE_DIR}/test/tablenames_test.cc
       ${CMAKE_SOURCE_DIR}/test/web_filesystem_test.cc
diff --git a/lib/src/webdb.cc b/lib/src/webdb.cc
index 0161443aa..3b9d1aa2e 100644
--- a/lib/src/webdb.cc
+++ b/lib/src/webdb.cc
@@ -301,6 +301,39 @@ arrow::Result<size_t> WebDB::Connection::CreatePreparedStatement(std::string_vie
     }
 }
 
+static arrow::Result<duckdb::vector<duckdb::Value>> BuildParameters(const rapidjson::Document::ConstArray& source) {
+    duckdb::vector<duckdb::Value> values;
+    size_t index = 0;
+
+    for (const auto& v : source) {
+        if (v.IsLosslessDouble())
+            values.emplace_back(v.GetDouble());
+        else if (v.IsString())
+            // Use GetStringLenght otherwise null bytes will be counted as terminators
+            values.emplace_back(string_t(v.GetString(), v.GetStringLength()));
+        else if (v.IsNull())
+            values.emplace_back(nullptr);
+        else if (v.IsBool())
+            values.emplace_back(v.GetBool());
+        else if (v.IsArray()) {
+            auto item_values = BuildParameters(v.GetArray());
+            if (!item_values.ok()) {
+                return item_values;
+            }
+            values.emplace_back(duckdb::Value::LIST(std::move(*item_values)));
+        } else
+            return arrow::Status{arrow::StatusCode::Invalid,
+                                 "Invalid column type encountered for argument " + std::to_string(index)};
+        ++index;
+    }
+
+    return arrow::Result<duckdb::vector<duckdb::Value>>(std::move(values));
+}
+
+static arrow::Result<duckdb::vector<duckdb::Value>> BuildParameters(const rapidjson::Document& doc) {
+    return BuildParameters(doc.GetArray());
+}
+
 arrow::Result<duckdb::unique_ptr<duckdb::QueryResult>> WebDB::Connection::ExecutePreparedStatement(
     size_t statement_id, std::string_view args_json) {
     try {
@@ -313,25 +346,12 @@ arrow::Result<duckdb::unique_ptr<duckdb::QueryResult>> WebDB::Connection::Execut
         if (!ok) return arrow::Status{arrow::StatusCode::Invalid, rapidjson::GetParseError_En(ok.Code())};
         if (!args_doc.IsArray()) return arrow::Status{arrow::StatusCode::Invalid, "Arguments must be given as array"};
 
-        duckdb::vector<duckdb::Value> values;
-        size_t index = 0;
-        for (const auto& v : args_doc.GetArray()) {
-            if (v.IsLosslessDouble())
-                values.emplace_back(v.GetDouble());
-            else if (v.IsString())
-                // Use GetStringLenght otherwise null bytes will be counted as terminators
-                values.emplace_back(string_t(v.GetString(), v.GetStringLength()));
-            else if (v.IsNull())
-                values.emplace_back(nullptr);
-            else if (v.IsBool())
-                values.emplace_back(v.GetBool());
-            else
-                return arrow::Status{arrow::StatusCode::Invalid,
-                                     "Invalid column type encountered for argument " + std::to_string(index)};
-            ++index;
+        auto values = BuildParameters(args_doc);
+        if (!values.ok()) {
+            return values.status();
         }
 
-        auto result = stmt->second->Execute(values);
+        auto result = stmt->second->Execute(*values);
         if (result->HasError()) return arrow::Status{arrow::StatusCode::ExecutionError, std::move(result->GetError())};
         return result;
     } catch (std::exception& e) {
diff --git a/lib/test/prepared_statement_test.cc b/lib/test/prepared_statement_test.cc
new file mode 100644
index 000000000..7e9ac5ec0
--- /dev/null
+++ b/lib/test/prepared_statement_test.cc
@@ -0,0 +1,74 @@
+#include <gmock/gmock.h>
+#include <gtest/gtest.h>
+
+#include <filesystem>
+
+#include "arrow/array/array_base.h"
+#include "arrow/array/array_primitive.h"
+#include "arrow/io/memory.h"
+#include "arrow/ipc/reader.h"
+#include "arrow/table.h"
+#include "duckdb/web/test/config.h"
+#include "duckdb/web/webdb.h"
+
+using namespace duckdb::web;
+using namespace std;
+namespace fs = std::filesystem;
+
+namespace {
+
+TEST(PreparedStatent, WithArrayParams) {
+    auto db = make_shared<WebDB>(NATIVE);
+    WebDB::Connection conn{*db};
+
+    auto data = test::SOURCE_DIR / ".." / "data" / "uni" / "studenten.parquet";
+    if (!fs::exists(data)) GTEST_SKIP_(": Missing data");
+
+    std::stringstream ss;
+    ss << "SELECT * FROM parquet_scan('" << data.string() << "') WHERE semester = ANY(?)";
+
+    auto stmt = conn.CreatePreparedStatement(ss.str());
+    ASSERT_TRUE(stmt.ok()) << stmt.status().message();
+
+    auto buffer = conn.RunPreparedStatement(*stmt, "[[12, 2]]");
+    ASSERT_TRUE(buffer.ok()) << buffer.status().message();
+
+    ::arrow::io::BufferReader buffer_reader(*buffer);
+    auto const reader = ::arrow::ipc::RecordBatchFileReader::Open(&buffer_reader);
+    ASSERT_TRUE(reader.ok()) << reader.status().message();
+
+    auto const batches = (*reader)->ToRecordBatches();
+    ASSERT_TRUE(batches.ok()) << batches.status().message();
+
+    for (auto& batch : *batches) {
+        auto const rows = batch->GetColumnByName("semester");
+        ASSERT_TRUE(rows) << "Must contain `semester` column";
+
+        auto const int_rows = dynamic_pointer_cast<arrow::Int32Array>(rows);
+
+        for (auto i = 0; i < int_rows->length(); ++i) {
+            EXPECT_THAT(int_rows->Value(i), testing::AnyOf(testing::Eq(2), testing::Eq(12)));
+        }
+    }
+}
+
+TEST(PreparedStatent, WithArrayParamsIllegal) {
+    auto db = make_shared<WebDB>(NATIVE);
+    WebDB::Connection conn{*db};
+
+    auto data = test::SOURCE_DIR / ".." / "data" / "uni" / "studenten.parquet";
+    if (!fs::exists(data)) GTEST_SKIP_(": Missing data");
+
+    std::stringstream ss;
+    ss << "SELECT * FROM parquet_scan('" << data.string() << "') WHERE semester = ANY(?)";
+
+    auto stmt = conn.CreatePreparedStatement(ss.str());
+    ASSERT_TRUE(stmt.ok()) << stmt.status().message();
+
+    // passed ununiformed type
+    auto buffer = conn.RunPreparedStatement(*stmt, "[[12, [2]]]");
+    ASSERT_FALSE(buffer.ok());
+    ASSERT_EQ(buffer.status().code(), arrow::StatusCode::ExecutionError);
+}
+
+}  // namespace