diff --git a/.github/workflows/diff.yaml b/.github/workflows/diff.yaml
index a7faa7d22..321937956 100644
--- a/.github/workflows/diff.yaml
+++ b/.github/workflows/diff.yaml
@@ -271,3 +271,30 @@ jobs:
           source ve3/bin/activate
           cd e2e
           LD_LIBRARY_PATH=$LD_LIBRARY_PATH:../../libs/mgclient/lib python runner.py --workloads-root-directory ./distributed_queries
+
+      - name: Run query performance tests
+        run: |
+          cd tests/manual
+          ./query_performance_runner.py
+
+      - name: Get branch name (merge)
+        if: github.event_name != 'pull_request'
+        shell: bash
+        run: echo "BRANCH_NAME=$(echo ${GITHUB_REF#refs/heads/} | tr / -)" >> $GITHUB_ENV
+
+      - name: Get branch name (pull request)
+        if: github.event_name == 'pull_request'
+        shell: bash
+        run: echo "BRANCH_NAME=$(echo ${GITHUB_HEAD_REF} | tr / -)" >> $GITHUB_ENV
+
+      - name: Upload macro benchmark results
+        run: |
+          cd tools/bench-graph-client
+          virtualenv -p python3 ve3
+          source ve3/bin/activate
+          pip install -r requirements.txt
+          ./main.py --benchmark-name "query_performance" \
+                    --benchmark-results-path "../../build/tests/manual/query_performance_benchmark/summary.json" \
+                    --github-run-id "${{ github.run_id }}" \
+                    --github-run-number "${{ github.run_number }}" \
+                    --head-branch-name "${{ env.BRANCH_NAME }}"
diff --git a/src/expr/ast/cypher_main_visitor.hpp b/src/expr/ast/cypher_main_visitor.hpp
index 17d2167b2..bbe5d2d06 100644
--- a/src/expr/ast/cypher_main_visitor.hpp
+++ b/src/expr/ast/cypher_main_visitor.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -51,7 +51,7 @@ constexpr char kId[] = "ID";
 
 namespace MG_INJECTED_NAMESPACE_NAME {
 namespace detail {
-using antlropencypher::MemgraphCypher;
+using antlropencypher::v2::MemgraphCypher;
 
 template <typename TVisitor>
 std::optional<std::pair<Expression *, size_t>> VisitMemoryLimit(MemgraphCypher::MemoryLimitContext *memory_limit_ctx,
@@ -211,13 +211,13 @@ inline std::string_view ToString(const PulsarConfigKey key) {
 }
 }  // namespace detail
 
-using antlropencypher::MemgraphCypher;
+using antlropencypher::v2::MemgraphCypher;
 
 struct ParsingContext {
   bool is_query_cached = false;
 };
 
-class CypherMainVisitor : public antlropencypher::MemgraphCypherBaseVisitor {
+class CypherMainVisitor : public antlropencypher::v2::MemgraphCypherBaseVisitor {
  public:
   explicit CypherMainVisitor(ParsingContext context, AstStorage *storage) : context_(context), storage_(storage) {}
 
diff --git a/src/io/future.hpp b/src/io/future.hpp
index 585f18938..99906f2de 100644
--- a/src/io/future.hpp
+++ b/src/io/future.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -120,7 +120,7 @@ class Shared {
       MG_ASSERT(!consumed_, "Promise filled after it was already consumed!");
       MG_ASSERT(!filled_, "Promise filled twice!");
 
-      item_ = item;
+      item_ = std::move(item);
       filled_ = true;
     }  // lock released before condition variable notification
 
@@ -235,7 +235,7 @@ class Promise {
   // Fill the expected item into the Future.
   void Fill(T item) {
     MG_ASSERT(!filled_or_moved_, "Promise::Fill called on a promise that is already filled or moved!");
-    shared_->Fill(item);
+    shared_->Fill(std::move(item));
     filled_or_moved_ = true;
   }
 
diff --git a/src/io/local_transport/local_transport.hpp b/src/io/local_transport/local_transport.hpp
index b64cabf1d..f17e5c059 100644
--- a/src/io/local_transport/local_transport.hpp
+++ b/src/io/local_transport/local_transport.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -30,10 +30,10 @@ class LocalTransport {
   explicit LocalTransport(std::shared_ptr<LocalTransportHandle> local_transport_handle)
       : local_transport_handle_(std::move(local_transport_handle)) {}
 
-  template <Message RequestT, Message ResponseT>
-  ResponseFuture<ResponseT> Request(Address to_address, Address from_address, RequestT request,
+  template <Message ResponseT, Message RequestT>
+  ResponseFuture<ResponseT> Request(Address to_address, Address from_address, RValueRef<RequestT> request,
                                     std::function<void()> fill_notifier, Duration timeout) {
-    return local_transport_handle_->template SubmitRequest<RequestT, ResponseT>(
+    return local_transport_handle_->template SubmitRequest<ResponseT, RequestT>(
         to_address, from_address, std::move(request), timeout, fill_notifier);
   }
 
@@ -43,8 +43,8 @@ class LocalTransport {
   }
 
   template <Message M>
-  void Send(Address to_address, Address from_address, RequestId request_id, M &&message) {
-    return local_transport_handle_->template Send<M>(to_address, from_address, request_id, std::forward<M>(message));
+  void Send(Address to_address, Address from_address, RequestId request_id, RValueRef<M> message) {
+    return local_transport_handle_->template Send<M>(to_address, from_address, request_id, std::move(message));
   }
 
   Time Now() const { return local_transport_handle_->Now(); }
diff --git a/src/io/local_transport/local_transport_handle.hpp b/src/io/local_transport/local_transport_handle.hpp
index 38538620f..cd19a3977 100644
--- a/src/io/local_transport/local_transport_handle.hpp
+++ b/src/io/local_transport/local_transport_handle.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -104,10 +104,10 @@ class LocalTransportHandle {
   }
 
   template <Message M>
-  void Send(Address to_address, Address from_address, RequestId request_id, M &&message) {
+  void Send(Address to_address, Address from_address, RequestId request_id, RValueRef<M> message) {
     auto type_info = TypeInfoFor(message);
 
-    std::any message_any(std::forward<M>(message));
+    std::any message_any(std::move(message));
     OpaqueMessage opaque_message{.to_address = to_address,
                                  .from_address = from_address,
                                  .request_id = request_id,
@@ -138,14 +138,14 @@ class LocalTransportHandle {
     cv_.notify_all();
   }
 
-  template <Message RequestT, Message ResponseT>
-  ResponseFuture<ResponseT> SubmitRequest(Address to_address, Address from_address, RequestT &&request,
+  template <Message ResponseT, Message RequestT>
+  ResponseFuture<ResponseT> SubmitRequest(Address to_address, Address from_address, RValueRef<RequestT> request,
                                           Duration timeout, std::function<void()> fill_notifier) {
     auto [future, promise] = memgraph::io::FuturePromisePairWithNotifications<ResponseResult<ResponseT>>(
         // set null notifier for when the Future::Wait is called
         nullptr,
         // set notifier for when Promise::Fill is called
-        std::forward<std::function<void()>>(fill_notifier));
+        std::move(fill_notifier));
 
     const bool port_matches = to_address.last_known_port == from_address.last_known_port;
     const bool ip_matches = to_address.last_known_ip == from_address.last_known_ip;
@@ -168,7 +168,7 @@ class LocalTransportHandle {
       promises_.emplace(std::move(promise_key), std::move(dop));
     }  // lock dropped
 
-    Send(to_address, from_address, request_id, std::forward<RequestT>(request));
+    Send<RequestT>(to_address, from_address, request_id, std::move(request));
 
     return std::move(future);
   }
diff --git a/src/io/rsm/raft.hpp b/src/io/rsm/raft.hpp
index 07a51288d..3be15cc46 100644
--- a/src/io/rsm/raft.hpp
+++ b/src/io/rsm/raft.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -19,6 +19,7 @@
 #include <map>
 #include <set>
 #include <thread>
+#include <type_traits>
 #include <vector>
 
 #include <boost/core/demangle.hpp>
@@ -246,7 +247,7 @@ to a CAS operation.
 template <typename WriteOperation, typename ReadOperation, typename ReplicatedState, typename WriteResponseValue,
           typename ReadResponseValue>
 concept Rsm = requires(ReplicatedState state, WriteOperation w, ReadOperation r) {
-  { state.Read(r) } -> std::same_as<ReadResponseValue>;
+  { state.Read(std::move(r)) } -> std::same_as<ReadResponseValue>;
   { state.Apply(w) } -> std::same_as<WriteResponseValue>;
 };
 
@@ -402,7 +403,7 @@ class Raft {
         const PendingClientRequest client_request = std::move(leader.pending_client_requests.at(apply_index));
         leader.pending_client_requests.erase(apply_index);
 
-        const WriteResponse<WriteResponseValue> resp{
+        WriteResponse<WriteResponseValue> resp{
             .success = true,
             .write_return = std::move(write_return),
             .raft_index = apply_index,
@@ -554,7 +555,7 @@ class Raft {
       for (const auto &peer : peers_) {
         // request_id not necessary to set because it's not a Future-backed Request.
         static constexpr auto request_id = 0;
-        io_.template Send<VoteRequest>(peer, request_id, request);
+        io_.template Send(peer, request_id, VoteRequest{request});
         outstanding_votes.insert(peer);
       }
 
@@ -624,13 +625,12 @@ class Raft {
       MG_ASSERT(std::max(req.term, state_.term) == req.term);
     }
 
-    const VoteResponse res{
-        .term = std::max(req.term, state_.term),
-        .committed_log_size = state_.committed_log_size,
-        .vote_granted = new_leader,
-    };
-
-    io_.Send(from_address, request_id, res);
+    io_.Send(from_address, request_id,
+             VoteResponse{
+                 .term = std::max(req.term, state_.term),
+                 .committed_log_size = state_.committed_log_size,
+                 .vote_granted = new_leader,
+             });
 
     if (new_leader) {
       // become a follower
@@ -718,6 +718,10 @@ class Raft {
         .log_size = state_.log.size(),
     };
 
+    static_assert(std::is_trivially_copyable_v<AppendResponse>,
+                  "This function copies this message, therefore it is important to be trivially copyable. Otherwise it "
+                  "should be moved");
+
     if constexpr (std::is_same<ALL, Leader>()) {
       MG_ASSERT(req.term != state_.term, "Multiple leaders are acting under the term ", req.term);
     }
@@ -736,7 +740,7 @@ class Raft {
       // become follower of this leader, reply with our log status
       state_.term = req.term;
 
-      io_.Send(from_address, request_id, res);
+      io_.Send(from_address, request_id, AppendResponse{res});
 
       Log("becoming Follower of Leader ", from_address.last_known_port, " at term ", req.term);
       return Follower{
@@ -747,7 +751,7 @@ class Raft {
 
     if (req.term < state_.term) {
       // nack this request from an old leader
-      io_.Send(from_address, request_id, res);
+      io_.Send(from_address, request_id, AppendResponse{res});
 
       return std::nullopt;
     }
@@ -808,7 +812,7 @@ class Raft {
 
     Log("returning log_size of ", res.log_size);
 
-    io_.Send(from_address, request_id, res);
+    io_.Send(from_address, request_id, AppendResponse{res});
 
     return std::nullopt;
   }
@@ -859,17 +863,17 @@ class Raft {
     auto type_info = TypeInfoFor(req);
     std::string demangled_name = boost::core::demangle(type_info.get().name());
     Log("handling ReadOperation<" + demangled_name + ">");
-    ReadOperation read_operation = req.operation;
+    ReadOperation &read_operation = req.operation;
 
-    ReadResponseValue read_return = replicated_state_.Read(read_operation);
+    ReadResponseValue read_return = replicated_state_.Read(std::move(read_operation));
 
-    const ReadResponse<ReadResponseValue> resp{
+    ReadResponse<ReadResponseValue> resp{
         .success = true,
         .read_return = std::move(read_return),
         .retry_leader = std::nullopt,
     };
 
-    io_.Send(from_address, request_id, resp);
+    io_.Send(from_address, request_id, std::move(resp));
 
     return std::nullopt;
   }
@@ -878,11 +882,11 @@ class Raft {
   std::optional<Role> Handle(Candidate & /* variable */, ReadRequest<ReadOperation> && /* variable */,
                              RequestId request_id, Address from_address) {
     Log("received ReadOperation - not redirecting because no Leader is known");
-    const ReadResponse<ReadResponseValue> res{
+    ReadResponse<ReadResponseValue> res{
         .success = false,
     };
 
-    io_.Send(from_address, request_id, res);
+    io_.Send(from_address, request_id, std::move(res));
 
     Cron();
 
@@ -894,12 +898,12 @@ class Raft {
                              Address from_address) {
     Log("redirecting client to known Leader with port ", follower.leader_address.last_known_port);
 
-    const ReadResponse<ReadResponseValue> res{
+    ReadResponse<ReadResponseValue> res{
         .success = false,
         .retry_leader = follower.leader_address,
     };
 
-    io_.Send(from_address, request_id, res);
+    io_.Send(from_address, request_id, std::move(res));
 
     return std::nullopt;
   }
@@ -913,12 +917,12 @@ class Raft {
                              Address from_address) {
     Log("redirecting client to known Leader with port ", follower.leader_address.last_known_port);
 
-    const WriteResponse<WriteResponseValue> res{
+    WriteResponse<WriteResponseValue> res{
         .success = false,
         .retry_leader = follower.leader_address,
     };
 
-    io_.Send(from_address, request_id, res);
+    io_.Send(from_address, request_id, std::move(res));
 
     return std::nullopt;
   }
@@ -927,11 +931,11 @@ class Raft {
                              RequestId request_id, Address from_address) {
     Log("received WriteRequest - not redirecting because no Leader is known");
 
-    const WriteResponse<WriteResponseValue> res{
+    WriteResponse<WriteResponseValue> res{
         .success = false,
     };
 
-    io_.Send(from_address, request_id, res);
+    io_.Send(from_address, request_id, std::move(res));
 
     Cron();
 
diff --git a/src/io/rsm/rsm_client.hpp b/src/io/rsm/rsm_client.hpp
index 2b76b6399..bc5dc6a47 100644
--- a/src/io/rsm/rsm_client.hpp
+++ b/src/io/rsm/rsm_client.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -95,7 +95,7 @@ class RsmClient {
   BasicResult<TimedOut, WriteResponseT> SendWriteRequest(WriteRequestT req) {
     Notifier notifier;
     const ReadinessToken readiness_token{0};
-    SendAsyncWriteRequest(req, notifier, readiness_token);
+    SendAsyncWriteRequest(std::move(req), notifier, readiness_token);
     auto poll_result = AwaitAsyncWriteRequest(readiness_token);
     while (!poll_result) {
       poll_result = AwaitAsyncWriteRequest(readiness_token);
@@ -106,7 +106,7 @@ class RsmClient {
   BasicResult<TimedOut, ReadResponseT> SendReadRequest(ReadRequestT req) {
     Notifier notifier;
     const ReadinessToken readiness_token{0};
-    SendAsyncReadRequest(req, notifier, readiness_token);
+    SendAsyncReadRequest(std::move(req), notifier, readiness_token);
     auto poll_result = AwaitAsyncReadRequest(readiness_token);
     while (!poll_result) {
       poll_result = AwaitAsyncReadRequest(readiness_token);
@@ -115,15 +115,15 @@ class RsmClient {
   }
 
   /// AsyncRead methods
-  void SendAsyncReadRequest(const ReadRequestT &req, Notifier notifier, ReadinessToken readiness_token) {
+  void SendAsyncReadRequest(ReadRequestT &&req, Notifier notifier, ReadinessToken readiness_token) {
     ReadRequest<ReadRequestT> read_req = {.operation = req};
 
     AsyncRequest<ReadRequestT, ReadResponse<ReadResponseT>> async_request{
         .start_time = io_.Now(),
         .request = std::move(req),
         .notifier = notifier,
-        .future = io_.template RequestWithNotification<ReadRequest<ReadRequestT>, ReadResponse<ReadResponseT>>(
-            leader_, read_req, notifier, readiness_token),
+        .future = io_.template RequestWithNotification<ReadResponse<ReadResponseT>, ReadRequest<ReadRequestT>>(
+            leader_, std::move(read_req), notifier, readiness_token),
     };
 
     async_reads_.emplace(readiness_token.GetId(), std::move(async_request));
@@ -134,8 +134,8 @@ class RsmClient {
 
     ReadRequest<ReadRequestT> read_req = {.operation = async_request.request};
 
-    async_request.future = io_.template RequestWithNotification<ReadRequest<ReadRequestT>, ReadResponse<ReadResponseT>>(
-        leader_, read_req, async_request.notifier, readiness_token);
+    async_request.future = io_.template RequestWithNotification<ReadResponse<ReadResponseT>, ReadRequest<ReadRequestT>>(
+        leader_, std::move(read_req), async_request.notifier, readiness_token);
   }
 
   std::optional<BasicResult<TimedOut, ReadResponseT>> PollAsyncReadRequest(const ReadinessToken &readiness_token) {
@@ -184,15 +184,15 @@ class RsmClient {
   }
 
   /// AsyncWrite methods
-  void SendAsyncWriteRequest(const WriteRequestT &req, Notifier notifier, ReadinessToken readiness_token) {
+  void SendAsyncWriteRequest(WriteRequestT &&req, Notifier notifier, ReadinessToken readiness_token) {
     WriteRequest<WriteRequestT> write_req = {.operation = req};
 
     AsyncRequest<WriteRequestT, WriteResponse<WriteResponseT>> async_request{
         .start_time = io_.Now(),
         .request = std::move(req),
         .notifier = notifier,
-        .future = io_.template RequestWithNotification<WriteRequest<WriteRequestT>, WriteResponse<WriteResponseT>>(
-            leader_, write_req, notifier, readiness_token),
+        .future = io_.template RequestWithNotification<WriteResponse<WriteResponseT>, WriteRequest<WriteRequestT>>(
+            leader_, std::move(write_req), notifier, readiness_token),
     };
 
     async_writes_.emplace(readiness_token.GetId(), std::move(async_request));
@@ -204,8 +204,8 @@ class RsmClient {
     WriteRequest<WriteRequestT> write_req = {.operation = async_request.request};
 
     async_request.future =
-        io_.template RequestWithNotification<WriteRequest<WriteRequestT>, WriteResponse<WriteResponseT>>(
-            leader_, write_req, async_request.notifier, readiness_token);
+        io_.template RequestWithNotification<WriteResponse<WriteResponseT>, WriteRequest<WriteRequestT>>(
+            leader_, std::move(write_req), async_request.notifier, readiness_token);
   }
 
   std::optional<BasicResult<TimedOut, WriteResponseT>> PollAsyncWriteRequest(const ReadinessToken &readiness_token) {
diff --git a/src/io/simulator/simulator_handle.hpp b/src/io/simulator/simulator_handle.hpp
index 3fd9b4965..a34e93c66 100644
--- a/src/io/simulator/simulator_handle.hpp
+++ b/src/io/simulator/simulator_handle.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -105,19 +105,19 @@ class SimulatorHandle {
 
   bool ShouldShutDown() const;
 
-  template <Message Request, Message Response>
-  ResponseFuture<Response> SubmitRequest(Address to_address, Address from_address, Request &&request, Duration timeout,
-                                         std::function<bool()> &&maybe_tick_simulator,
-                                         std::function<void()> &&fill_notifier) {
+  template <Message ResponseT, Message RequestT>
+  ResponseFuture<ResponseT> SubmitRequest(Address to_address, Address from_address, RValueRef<RequestT> request,
+                                          Duration timeout, std::function<bool()> &&maybe_tick_simulator,
+                                          std::function<void()> &&fill_notifier) {
     auto type_info = TypeInfoFor(request);
     std::string demangled_name = boost::core::demangle(type_info.get().name());
     spdlog::trace("simulator sending request {} to {}", demangled_name, to_address);
 
-    auto [future, promise] = memgraph::io::FuturePromisePairWithNotifications<ResponseResult<Response>>(
+    auto [future, promise] = memgraph::io::FuturePromisePairWithNotifications<ResponseResult<ResponseT>>(
         // set notifier for when the Future::Wait is called
-        std::forward<std::function<bool()>>(maybe_tick_simulator),
+        std::move(maybe_tick_simulator),
         // set notifier for when Promise::Fill is called
-        std::forward<std::function<void()>>(fill_notifier));
+        std::move(fill_notifier));
 
     {
       std::unique_lock<std::mutex> lock(mu_);
@@ -126,7 +126,7 @@ class SimulatorHandle {
 
       const Time deadline = cluster_wide_time_microseconds_ + timeout;
 
-      std::any message(request);
+      std::any message(std::move(request));
       OpaqueMessage om{.to_address = to_address,
                        .from_address = from_address,
                        .request_id = request_id,
@@ -194,7 +194,7 @@ class SimulatorHandle {
   }
 
   template <Message M>
-  void Send(Address to_address, Address from_address, RequestId request_id, M message) {
+  void Send(Address to_address, Address from_address, RequestId request_id, RValueRef<M> message) {
     spdlog::trace("sending message from {} to {}", from_address.last_known_port, to_address.last_known_port);
     auto type_info = TypeInfoFor(message);
     {
diff --git a/src/io/simulator/simulator_transport.hpp b/src/io/simulator/simulator_transport.hpp
index 1272a04a1..05775e726 100644
--- a/src/io/simulator/simulator_transport.hpp
+++ b/src/io/simulator/simulator_transport.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -33,14 +33,14 @@ class SimulatorTransport {
   SimulatorTransport(std::shared_ptr<SimulatorHandle> simulator_handle, Address address, uint64_t seed)
       : simulator_handle_(simulator_handle), address_(address), rng_(std::mt19937{seed}) {}
 
-  template <Message RequestT, Message ResponseT>
-  ResponseFuture<ResponseT> Request(Address to_address, Address from_address, RequestT request,
+  template <Message ResponseT, Message RequestT>
+  ResponseFuture<ResponseT> Request(Address to_address, Address from_address, RValueRef<RequestT> request,
                                     std::function<void()> notification, Duration timeout) {
     std::function<bool()> tick_simulator = [handle_copy = simulator_handle_] {
       return handle_copy->MaybeTickSimulator();
     };
 
-    return simulator_handle_->template SubmitRequest<RequestT, ResponseT>(
+    return simulator_handle_->template SubmitRequest<ResponseT, RequestT>(
         to_address, from_address, std::move(request), timeout, std::move(tick_simulator), std::move(notification));
   }
 
@@ -50,8 +50,8 @@ class SimulatorTransport {
   }
 
   template <Message M>
-  void Send(Address to_address, Address from_address, uint64_t request_id, M message) {
-    return simulator_handle_->template Send<M>(to_address, from_address, request_id, message);
+  void Send(Address to_address, Address from_address, uint64_t request_id, RValueRef<M> message) {
+    return simulator_handle_->template Send<M>(to_address, from_address, request_id, std::move(message));
   }
 
   Time Now() const { return simulator_handle_->Now(); }
diff --git a/src/io/transport.hpp b/src/io/transport.hpp
index 5dd7a9a39..eade62db3 100644
--- a/src/io/transport.hpp
+++ b/src/io/transport.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -22,6 +22,7 @@
 #include "io/message_histogram_collector.hpp"
 #include "io/notifier.hpp"
 #include "io/time.hpp"
+#include "utils/concepts.hpp"
 #include "utils/result.hpp"
 
 namespace memgraph::io {
@@ -32,7 +33,15 @@ using memgraph::utils::BasicResult;
 // reasonable constraints around message types over time,
 // as we adapt things to use Thrift-generated message types.
 template <typename T>
-concept Message = std::same_as<T, std::decay_t<T>>;
+concept Message = std::movable<T> && std::copyable<T>;
+
+template <utils::Object T>
+struct RValueRefEnforcer {
+  using Type = T &&;
+};
+
+template <typename T>
+using RValueRef = typename RValueRefEnforcer<T>::Type;
 
 using RequestId = uint64_t;
 
@@ -82,44 +91,44 @@ class Io {
   Duration GetDefaultTimeout() { return default_timeout_; }
 
   /// Issue a request with an explicit timeout in microseconds provided. This tends to be used by clients.
-  template <Message RequestT, Message ResponseT>
-  ResponseFuture<ResponseT> RequestWithTimeout(Address address, RequestT request, Duration timeout) {
+  template <Message ResponseT, Message RequestT>
+  ResponseFuture<ResponseT> RequestWithTimeout(Address address, RValueRef<RequestT> request, Duration timeout) {
     const Address from_address = address_;
     std::function<void()> fill_notifier = nullptr;
-    return implementation_.template Request<RequestT, ResponseT>(address, from_address, request, fill_notifier,
-                                                                 timeout);
+    return implementation_.template Request<ResponseT, RequestT>(address, from_address, std::move(request),
+                                                                 fill_notifier, timeout);
   }
 
   /// Issue a request that times out after the default timeout. This tends
   /// to be used by clients.
-  template <Message RequestT, Message ResponseT>
-  ResponseFuture<ResponseT> Request(Address to_address, RequestT request) {
+  template <Message ResponseT, Message RequestT>
+  ResponseFuture<ResponseT> Request(Address to_address, RValueRef<RequestT> request) {
     const Duration timeout = default_timeout_;
     const Address from_address = address_;
     std::function<void()> fill_notifier = nullptr;
-    return implementation_.template Request<RequestT, ResponseT>(to_address, from_address, std::move(request),
+    return implementation_.template Request<ResponseT, RequestT>(to_address, from_address, std::move(request),
                                                                  fill_notifier, timeout);
   }
 
   /// Issue a request that will notify a Notifier when it is filled or times out.
-  template <Message RequestT, Message ResponseT>
-  ResponseFuture<ResponseT> RequestWithNotification(Address to_address, RequestT request, Notifier notifier,
+  template <Message ResponseT, Message RequestT>
+  ResponseFuture<ResponseT> RequestWithNotification(Address to_address, RValueRef<RequestT> request, Notifier notifier,
                                                     ReadinessToken readiness_token) {
     const Duration timeout = default_timeout_;
     const Address from_address = address_;
     std::function<void()> fill_notifier = [notifier, readiness_token]() { notifier.Notify(readiness_token); };
-    return implementation_.template Request<RequestT, ResponseT>(to_address, from_address, std::move(request),
+    return implementation_.template Request<ResponseT, RequestT>(to_address, from_address, std::move(request),
                                                                  fill_notifier, timeout);
   }
 
   /// Issue a request that will notify a Notifier when it is filled or times out.
-  template <Message RequestT, Message ResponseT>
-  ResponseFuture<ResponseT> RequestWithNotificationAndTimeout(Address to_address, RequestT request, Notifier notifier,
+  template <Message ResponseT, Message RequestT>
+  ResponseFuture<ResponseT> RequestWithNotificationAndTimeout(Address to_address, RequestT &&request, Notifier notifier,
                                                               ReadinessToken readiness_token, Duration timeout) {
     const Address from_address = address_;
     std::function<void()> fill_notifier = [notifier, readiness_token]() { notifier.Notify(readiness_token); };
-    return implementation_.template Request<RequestT, ResponseT>(to_address, from_address, std::move(request),
-                                                                 fill_notifier, timeout);
+    return implementation_.template Request<ResponseT>(to_address, from_address, std::forward<RequestT>(request),
+                                                       fill_notifier, timeout);
   }
 
   /// Wait for an explicit number of microseconds for a request of one of the
@@ -141,9 +150,9 @@ class Io {
   /// responses are not necessarily expected, and for servers to respond to requests.
   /// If you need reliable delivery, this must be built on-top. TCP is not enough for most use cases.
   template <Message M>
-  void Send(Address to_address, RequestId request_id, M message) {
+  void Send(Address to_address, RequestId request_id, M &&message) {
     Address from_address = address_;
-    return implementation_.template Send<M>(to_address, from_address, request_id, std::move(message));
+    return implementation_.template Send<M>(to_address, from_address, request_id, std::forward<M>(message));
   }
 
   /// The current system time. This time source should be preferred over any other,
diff --git a/src/parser/CMakeLists.txt b/src/parser/CMakeLists.txt
index 7575b0529..b7ae43178 100644
--- a/src/parser/CMakeLists.txt
+++ b/src/parser/CMakeLists.txt
@@ -23,7 +23,7 @@ add_custom_command(
     COMMAND ${CMAKE_COMMAND} -E make_directory ${opencypher_generated}
     COMMAND
         java -jar ${CMAKE_SOURCE_DIR}/libs/antlr-4.10.1-complete.jar
-            -Dlanguage=Cpp -visitor -package antlropencypher
+            -Dlanguage=Cpp -visitor -package antlropencypher::v2
             -o ${opencypher_generated}
             ${opencypher_lexer_grammar} ${opencypher_parser_grammar}
     WORKING_DIRECTORY "${CMAKE_BINARY_DIR}"
diff --git a/src/parser/opencypher/parser.hpp b/src/parser/opencypher/parser.hpp
index 9a57bc65b..25fdabf4b 100644
--- a/src/parser/opencypher/parser.hpp
+++ b/src/parser/opencypher/parser.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -14,10 +14,10 @@
 #include <string>
 
 #include "antlr4-runtime.h"
-#include "utils/exceptions.hpp"
 #include "parser/opencypher/generated/MemgraphCypher.h"
 #include "parser/opencypher/generated/MemgraphCypherLexer.h"
 #include "utils/concepts.hpp"
+#include "utils/exceptions.hpp"
 
 namespace memgraph::frontend::opencypher {
 
@@ -32,11 +32,9 @@ class SyntaxException : public utils::BasicException {
  * This thing must me a class since parser.cypher() returns pointer and there is
  * no way for us to get ownership over the object.
  */
-enum class ParserOpTag : uint8_t {
-  CYPHER, EXPRESSION
-};
+enum class ParserOpTag : uint8_t { CYPHER, EXPRESSION };
 
-template<ParserOpTag Tag = ParserOpTag::CYPHER>
+template <ParserOpTag Tag = ParserOpTag::CYPHER>
 class Parser {
  public:
   /**
@@ -46,10 +44,9 @@ class Parser {
   Parser(const std::string query) : query_(std::move(query)) {
     parser_.removeErrorListeners();
     parser_.addErrorListener(&error_listener_);
-    if constexpr(Tag == ParserOpTag::CYPHER) {
+    if constexpr (Tag == ParserOpTag::CYPHER) {
       tree_ = parser_.cypher();
-    }
-    else {
+    } else {
       tree_ = parser_.expression();
     }
     if (parser_.getNumberOfSyntaxErrors()) {
@@ -75,11 +72,11 @@ class Parser {
   FirstMessageErrorListener error_listener_;
   std::string query_;
   antlr4::ANTLRInputStream input_{query_};
-  antlropencypher::MemgraphCypherLexer lexer_{&input_};
+  antlropencypher::v2::MemgraphCypherLexer lexer_{&input_};
   antlr4::CommonTokenStream tokens_{&lexer_};
 
   // generate ast
-  antlropencypher::MemgraphCypher parser_{&tokens_};
+  antlropencypher::v2::MemgraphCypher parser_{&tokens_};
   antlr4::tree::ParseTree *tree_ = nullptr;
 };
 }  // namespace memgraph::frontend::opencypher
diff --git a/src/query/CMakeLists.txt b/src/query/CMakeLists.txt
index 0303f2fa5..d5545cd95 100644
--- a/src/query/CMakeLists.txt
+++ b/src/query/CMakeLists.txt
@@ -48,18 +48,20 @@ add_dependencies(mg-query generate_lcp_query)
 target_include_directories(mg-query PUBLIC ${CMAKE_SOURCE_DIR}/include)
 target_link_libraries(mg-query dl cppitertools Boost::headers)
 target_link_libraries(mg-query mg-integrations-pulsar mg-integrations-kafka mg-storage-v2 mg-license mg-utils mg-kvstore mg-memory)
+
 if(NOT "${MG_PYTHON_PATH}" STREQUAL "")
     set(Python3_ROOT_DIR "${MG_PYTHON_PATH}")
 endif()
+
 if("${MG_PYTHON_VERSION}" STREQUAL "")
     find_package(Python3 3.5 REQUIRED COMPONENTS Development)
 else()
     find_package(Python3 "${MG_PYTHON_VERSION}" EXACT REQUIRED COMPONENTS Development)
 endif()
+
 target_link_libraries(mg-query Python3::Python)
 
 # Generate Antlr openCypher parser
-
 set(opencypher_frontend ${CMAKE_CURRENT_SOURCE_DIR}/frontend/opencypher)
 set(opencypher_generated ${opencypher_frontend}/generated)
 set(opencypher_lexer_grammar ${opencypher_frontend}/grammar/MemgraphCypherLexer.g4)
@@ -82,15 +84,15 @@ add_custom_command(
     OUTPUT ${antlr_opencypher_generated_src} ${antlr_opencypher_generated_include}
     COMMAND ${CMAKE_COMMAND} -E make_directory ${opencypher_generated}
     COMMAND
-        java -jar ${CMAKE_SOURCE_DIR}/libs/antlr-4.10.1-complete.jar
-            -Dlanguage=Cpp -visitor -package antlropencypher
-            -o ${opencypher_generated}
-            ${opencypher_lexer_grammar} ${opencypher_parser_grammar}
+    java -jar ${CMAKE_SOURCE_DIR}/libs/antlr-4.10.1-complete.jar
+    -Dlanguage=Cpp -visitor -package antlropencypher
+    -o ${opencypher_generated}
+    ${opencypher_lexer_grammar} ${opencypher_parser_grammar}
     WORKING_DIRECTORY "${CMAKE_BINARY_DIR}"
     DEPENDS
-        ${opencypher_lexer_grammar} ${opencypher_parser_grammar}
-        ${opencypher_frontend}/grammar/CypherLexer.g4
-        ${opencypher_frontend}/grammar/Cypher.g4)
+    ${opencypher_lexer_grammar} ${opencypher_parser_grammar}
+    ${opencypher_frontend}/grammar/CypherLexer.g4
+    ${opencypher_frontend}/grammar/Cypher.g4)
 
 add_custom_target(generate_opencypher_parser
     DEPENDS ${antlr_opencypher_generated_src} ${antlr_opencypher_generated_include})
diff --git a/src/query/v2/cypher_query_interpreter.cpp b/src/query/v2/cypher_query_interpreter.cpp
index f3f8e48d7..6ac6b36fa 100644
--- a/src/query/v2/cypher_query_interpreter.cpp
+++ b/src/query/v2/cypher_query_interpreter.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -14,9 +14,9 @@
 #include "query/v2/request_router.hpp"
 
 // NOLINTNEXTLINE (cppcoreguidelines-avoid-non-const-global-variables)
-DEFINE_HIDDEN_bool(query_cost_planner, true, "Use the cost-estimating query planner.");
+DEFINE_HIDDEN_bool(query_v2_cost_planner, true, "Use the cost-estimating query planner.");
 // NOLINTNEXTLINE (cppcoreguidelines-avoid-non-const-global-variables)
-DEFINE_VALIDATED_int32(query_plan_cache_ttl, 60, "Time to live for cached query plans, in seconds.",
+DEFINE_VALIDATED_int32(query_v2_plan_cache_ttl, 60, "Time to live for cached query plans, in seconds.",
                        FLAG_IN_RANGE(0, std::numeric_limits<int32_t>::max()));
 
 namespace memgraph::query::v2 {
@@ -123,7 +123,7 @@ std::unique_ptr<LogicalPlan> MakeLogicalPlan(AstStorage ast_storage, CypherQuery
   auto vertex_counts = plan::MakeVertexCountCache(request_router);
   auto symbol_table = expr::MakeSymbolTable(query, predefined_identifiers);
   auto planning_context = plan::MakePlanningContext(&ast_storage, &symbol_table, query, &vertex_counts);
-  auto [root, cost] = plan::MakeLogicalPlan(&planning_context, parameters, FLAGS_query_cost_planner);
+  auto [root, cost] = plan::MakeLogicalPlan(&planning_context, parameters, FLAGS_query_v2_cost_planner);
   return std::make_unique<SingleNodeLogicalPlan>(std::move(root), cost, std::move(ast_storage),
                                                  std::move(symbol_table));
 }
diff --git a/src/query/v2/cypher_query_interpreter.hpp b/src/query/v2/cypher_query_interpreter.hpp
index 688e52fed..18505820f 100644
--- a/src/query/v2/cypher_query_interpreter.hpp
+++ b/src/query/v2/cypher_query_interpreter.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -22,9 +22,9 @@
 #include "utils/timer.hpp"
 
 // NOLINTNEXTLINE (cppcoreguidelines-avoid-non-const-global-variables)
-DECLARE_bool(query_cost_planner);
+DECLARE_bool(query_v2_cost_planner);
 // NOLINTNEXTLINE (cppcoreguidelines-avoid-non-const-global-variables)
-DECLARE_int32(query_plan_cache_ttl);
+DECLARE_int32(query_v2_plan_cache_ttl);
 
 namespace memgraph::query::v2 {
 
@@ -58,7 +58,7 @@ class CachedPlan {
 
   bool IsExpired() const {
     // NOLINTNEXTLINE (modernize-use-nullptr)
-    return cache_timer_.Elapsed() > std::chrono::seconds(FLAGS_query_plan_cache_ttl);
+    return cache_timer_.Elapsed() > std::chrono::seconds(FLAGS_query_v2_plan_cache_ttl);
   };
 
  private:
diff --git a/src/query/v2/plan/operator.cpp b/src/query/v2/plan/operator.cpp
index 993d24282..39ccf51ef 100644
--- a/src/query/v2/plan/operator.cpp
+++ b/src/query/v2/plan/operator.cpp
@@ -3087,25 +3087,18 @@ class DistributedExpandCursor : public Cursor {
     MG_ASSERT(direction != EdgeAtom::Direction::BOTH);
     const auto &edge = frame[self_.common_.edge_symbol].ValueEdge();
     static constexpr auto get_dst_vertex = [](const EdgeAccessor &edge,
-                                              const EdgeAtom::Direction direction) -> msgs::VertexId {
+                                              const EdgeAtom::Direction direction) -> accessors::VertexAccessor {
       switch (direction) {
         case EdgeAtom::Direction::IN:
-          return edge.From().Id();
+          return edge.From();
         case EdgeAtom::Direction::OUT:
-          return edge.To().Id();
+          return edge.To();
         case EdgeAtom::Direction::BOTH:
           throw std::runtime_error("EdgeDirection Both not implemented");
       }
     };
 
-    msgs::GetPropertiesRequest request;
-    // to not fetch any properties of the edges
-    request.vertex_ids.push_back(get_dst_vertex(edge, direction));
-    auto result_rows = context.request_router->GetProperties(std::move(request));
-    MG_ASSERT(result_rows.size() == 1);
-    auto &result_row = result_rows.front();
-    frame[self_.common_.node_symbol] =
-        accessors::VertexAccessor(msgs::Vertex{result_row.vertex}, result_row.props, context.request_router);
+    frame[self_.common_.node_symbol] = get_dst_vertex(edge, direction);
   }
 
   bool InitEdges(Frame &frame, ExecutionContext &context) {
@@ -3129,6 +3122,8 @@ class DistributedExpandCursor : public Cursor {
       // to not fetch any properties of the edges
       request.edge_properties.emplace();
       request.src_vertices.push_back(vertex.Id());
+      request.edge_properties.emplace();
+      request.src_vertex_properties.emplace();
       auto result_rows = std::invoke([&context, &request]() mutable {
         SCOPED_REQUEST_WAIT_PROFILE;
         return context.request_router->ExpandOne(std::move(request));
@@ -3271,6 +3266,7 @@ class DistributedExpandCursor : public Cursor {
                    [](const storage::v3::EdgeTypeId edge_type_id) { return msgs::EdgeType{edge_type_id}; });
     // to not fetch any properties of the edges
     request.edge_properties.emplace();
+    request.src_vertex_properties.emplace();
     for (const auto &frame : own_multi_frame_->GetValidFramesReader()) {
       const auto &vertex_value = frame[self_.input_symbol_];
 
diff --git a/src/query/v2/plan/rewrite/index_lookup.cpp b/src/query/v2/plan/rewrite/index_lookup.cpp
index 795f15fa4..b12a26da1 100644
--- a/src/query/v2/plan/rewrite/index_lookup.cpp
+++ b/src/query/v2/plan/rewrite/index_lookup.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -13,7 +13,8 @@
 
 #include "utils/flag_validation.hpp"
 
-DEFINE_VALIDATED_HIDDEN_int64(query_vertex_count_to_expand_existing, 10,
+// NOLINTNEXTLINE (cppcoreguidelines-avoid-non-const-global-variables)
+DEFINE_VALIDATED_HIDDEN_int64(query_v2_vertex_count_to_expand_existing, 10,
                               "Maximum count of indexed vertices which provoke "
                               "indexed lookup and then expand to existing, instead of "
                               "a regular expand. Default is 10, to turn off use -1.",
diff --git a/src/query/v2/plan/rewrite/index_lookup.hpp b/src/query/v2/plan/rewrite/index_lookup.hpp
index 0b9b9cb97..7f5b601e5 100644
--- a/src/query/v2/plan/rewrite/index_lookup.hpp
+++ b/src/query/v2/plan/rewrite/index_lookup.hpp
@@ -30,7 +30,8 @@
 #include "query/v2/plan/preprocess.hpp"
 #include "storage/v3/id_types.hpp"
 
-DECLARE_int64(query_vertex_count_to_expand_existing);
+// NOLINTNEXTLINE (cppcoreguidelines-avoid-non-const-global-variables)
+DECLARE_int64(query_v2_vertex_count_to_expand_existing);
 
 namespace memgraph::query::v2::plan {
 
@@ -100,7 +101,7 @@ class IndexLookupRewriter final : public HierarchicalLogicalOperatorVisitor {
       return true;
     }
     ScanAll dst_scan(expand.input(), expand.common_.node_symbol, expand.view_);
-    auto indexed_scan = GenScanByIndex(dst_scan, FLAGS_query_vertex_count_to_expand_existing);
+    auto indexed_scan = GenScanByIndex(dst_scan, FLAGS_query_v2_vertex_count_to_expand_existing);
     if (indexed_scan) {
       expand.set_input(std::move(indexed_scan));
       expand.common_.existing_node = true;
@@ -129,7 +130,7 @@ class IndexLookupRewriter final : public HierarchicalLogicalOperatorVisitor {
       // unconditionally creating an indexed scan.
       indexed_scan = GenScanByIndex(dst_scan);
     } else {
-      indexed_scan = GenScanByIndex(dst_scan, FLAGS_query_vertex_count_to_expand_existing);
+      indexed_scan = GenScanByIndex(dst_scan, FLAGS_query_v2_vertex_count_to_expand_existing);
     }
     if (indexed_scan) {
       expand.set_input(std::move(indexed_scan));
diff --git a/src/query/v2/plan/variable_start_planner.cpp b/src/query/v2/plan/variable_start_planner.cpp
index b6d15f73d..ec23be1af 100644
--- a/src/query/v2/plan/variable_start_planner.cpp
+++ b/src/query/v2/plan/variable_start_planner.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -17,7 +17,8 @@
 #include "utils/flag_validation.hpp"
 #include "utils/logging.hpp"
 
-DEFINE_VALIDATED_HIDDEN_uint64(query_max_plans, 1000U, "Maximum number of generated plans for a query.",
+// NOLINTNEXTLINE (cppcoreguidelines-avoid-non-const-global-variables)
+DEFINE_VALIDATED_HIDDEN_uint64(query_v2_max_plans, 1000U, "Maximum number of generated plans for a query.",
                                FLAG_IN_RANGE(1, std::numeric_limits<std::uint64_t>::max()));
 
 namespace memgraph::query::v2::plan::impl {
diff --git a/src/query/v2/plan/variable_start_planner.hpp b/src/query/v2/plan/variable_start_planner.hpp
index 27722b6b2..01d7cef58 100644
--- a/src/query/v2/plan/variable_start_planner.hpp
+++ b/src/query/v2/plan/variable_start_planner.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -18,7 +18,8 @@
 
 #include "query/v2/plan/rule_based_planner.hpp"
 
-DECLARE_uint64(query_max_plans);
+// NOLINTNEXTLINE (cppcoreguidelines-avoid-non-const-global-variables)
+DECLARE_uint64(query_v2_max_plans);
 
 namespace memgraph::query::v2::plan {
 
@@ -310,7 +311,7 @@ class VariableStartPlanner {
     for (const auto &query_part : query_parts) {
       alternative_query_parts.emplace_back(impl::VaryQueryPartMatching(query_part, symbol_table));
     }
-    return iter::slice(MakeCartesianProduct(std::move(alternative_query_parts)), 0UL, FLAGS_query_max_plans);
+    return iter::slice(MakeCartesianProduct(std::move(alternative_query_parts)), 0UL, FLAGS_query_v2_max_plans);
   }
 
  public:
diff --git a/src/query/v2/request_router.hpp b/src/query/v2/request_router.hpp
index a8326d900..04107a73e 100644
--- a/src/query/v2/request_router.hpp
+++ b/src/query/v2/request_router.hpp
@@ -323,8 +323,8 @@ class RequestRouter : public RequestRouterInterface {
       io::ReadinessToken readiness_token{i};
       auto &storage_client = GetStorageClientForShard(request.shard);
       msgs::WriteRequests req = request.request;
-      storage_client.SendAsyncWriteRequest(req, notifier_, readiness_token);
-      running_requests.emplace(readiness_token.GetId(), request);
+      storage_client.SendAsyncWriteRequest(std::move(req), notifier_, readiness_token);
+      running_requests.emplace(readiness_token.GetId(), std::move(request));
     }
 
     // drive requests to completion
@@ -339,7 +339,8 @@ class RequestRouter : public RequestRouterInterface {
     // must be fetched again with an ExpandOne(Edges.dst)
 
     // create requests
-    std::vector<ShardRequestState<msgs::ExpandOneRequest>> requests_to_be_sent = RequestsForExpandOne(request);
+    std::vector<ShardRequestState<msgs::ExpandOneRequest>> requests_to_be_sent =
+        RequestsForExpandOne(std::move(request));
 
     // begin all requests in parallel
     RunningRequests<msgs::ExpandOneRequest> running_requests = {};
@@ -349,8 +350,8 @@ class RequestRouter : public RequestRouterInterface {
       io::ReadinessToken readiness_token{i};
       auto &storage_client = GetStorageClientForShard(request.shard);
       msgs::ReadRequests req = request.request;
-      storage_client.SendAsyncReadRequest(req, notifier_, readiness_token);
-      running_requests.emplace(readiness_token.GetId(), request);
+      storage_client.SendAsyncReadRequest(std::move(req), notifier_, readiness_token);
+      running_requests.emplace(readiness_token.GetId(), std::move(request));
     }
 
     // drive requests to completion
@@ -386,8 +387,8 @@ class RequestRouter : public RequestRouterInterface {
       io::ReadinessToken readiness_token{i};
       auto &storage_client = GetStorageClientForShard(request.shard);
       msgs::ReadRequests req = request.request;
-      storage_client.SendAsyncReadRequest(req, notifier_, readiness_token);
-      running_requests.emplace(readiness_token.GetId(), request);
+      storage_client.SendAsyncReadRequest(std::move(req), notifier_, readiness_token);
+      running_requests.emplace(readiness_token.GetId(), std::move(request));
     }
 
     // drive requests to completion
@@ -503,6 +504,7 @@ class RequestRouter : public RequestRouterInterface {
 
         msgs::ScanVerticesRequest request;
         request.transaction_id = transaction_id_;
+        request.props_to_return.emplace();
         request.start_id.second = storage::conversions::ConvertValueVector(key);
 
         ShardRequestState<msgs::ScanVerticesRequest> shard_request_state{
@@ -517,7 +519,7 @@ class RequestRouter : public RequestRouterInterface {
     return requests;
   }
 
-  std::vector<ShardRequestState<msgs::ExpandOneRequest>> RequestsForExpandOne(const msgs::ExpandOneRequest &request) {
+  std::vector<ShardRequestState<msgs::ExpandOneRequest>> RequestsForExpandOne(msgs::ExpandOneRequest &&request) {
     std::map<ShardMetadata, msgs::ExpandOneRequest> per_shard_request_table;
     msgs::ExpandOneRequest top_level_rqst_template = request;
     top_level_rqst_template.transaction_id = transaction_id_;
@@ -529,7 +531,7 @@ class RequestRouter : public RequestRouterInterface {
       if (!per_shard_request_table.contains(shard)) {
         per_shard_request_table.insert(std::pair(shard, top_level_rqst_template));
       }
-      per_shard_request_table[shard].src_vertices.push_back(vertex);
+      per_shard_request_table[shard].src_vertices.push_back(std::move(vertex));
     }
 
     std::vector<ShardRequestState<msgs::ExpandOneRequest>> requests = {};
@@ -726,11 +728,11 @@ class RequestRouter : public RequestRouterInterface {
     coordinator::CoordinatorWriteRequests requests{coordinator::AllocateEdgeIdBatchRequest{.batch_size = 1000000}};
 
     io::rsm::WriteRequest<coordinator::CoordinatorWriteRequests> ww;
-    ww.operation = requests;
-    auto resp =
-        io_.template Request<io::rsm::WriteRequest<coordinator::CoordinatorWriteRequests>,
-                             io::rsm::WriteResponse<coordinator::CoordinatorWriteResponses>>(coordinator_address, ww)
-            .Wait();
+    ww.operation = std::move(requests);
+    auto resp = io_.template Request<io::rsm::WriteResponse<coordinator::CoordinatorWriteResponses>,
+                                     io::rsm::WriteRequest<coordinator::CoordinatorWriteRequests>>(coordinator_address,
+                                                                                                   std::move(ww))
+                    .Wait();
     if (resp.HasValue()) {
       const auto alloc_edge_id_reps =
           std::get<coordinator::AllocateEdgeIdBatchResponse>(resp.GetValue().message.write_return);
diff --git a/src/storage/v3/expr.cpp b/src/storage/v3/expr.cpp
index 36c0ff8e3..185e4e29e 100644
--- a/src/storage/v3/expr.cpp
+++ b/src/storage/v3/expr.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -35,7 +35,7 @@ msgs::Value ConstructValueVertex(const VertexAccessor &acc, View view) {
   memgraph::msgs::Label value_label{.id = prim_label};
 
   auto prim_key = conversions::ConvertValueVector(acc.PrimaryKey(view).GetValue());
-  memgraph::msgs::VertexId vertex_id = std::make_pair(value_label, prim_key);
+  memgraph::msgs::VertexId vertex_id = std::make_pair(value_label, std::move(prim_key));
 
   // Get the labels
   auto vertex_labels = acc.Labels(view).GetValue();
@@ -45,7 +45,7 @@ msgs::Value ConstructValueVertex(const VertexAccessor &acc, View view) {
   std::transform(vertex_labels.begin(), vertex_labels.end(), std::back_inserter(value_labels),
                  [](const auto &label) { return msgs::Label{.id = label}; });
 
-  return msgs::Value({.id = vertex_id, .labels = value_labels});
+  return msgs::Value({.id = std::move(vertex_id), .labels = std::move(value_labels)});
 }
 
 msgs::Value ConstructValueEdge(const EdgeAccessor &acc, View view) {
diff --git a/src/storage/v3/request_helper.cpp b/src/storage/v3/request_helper.cpp
index f13c5a82e..8e20d8e58 100644
--- a/src/storage/v3/request_helper.cpp
+++ b/src/storage/v3/request_helper.cpp
@@ -46,7 +46,6 @@ struct VertexIdCmpr {
 std::optional<std::map<PropertyId, Value>> PrimaryKeysFromAccessor(const VertexAccessor &acc, View view,
                                                                    const Schemas::Schema &schema) {
   std::map<PropertyId, Value> ret;
-  auto props = acc.Properties(view);
   auto maybe_pk = acc.PrimaryKey(view);
   if (maybe_pk.HasError()) {
     spdlog::debug("Encountered an error while trying to get vertex primary key.");
@@ -58,7 +57,7 @@ std::optional<std::map<PropertyId, Value>> PrimaryKeysFromAccessor(const VertexA
     ret.emplace(schema.second[i].property_id, FromPropertyValueToValue(std::move(pk[i])));
   }
 
-  return ret;
+  return {std::move(ret)};
 }
 
 ShardResult<std::vector<msgs::Label>> FillUpSourceVertexSecondaryLabels(const std::optional<VertexAccessor> &v_acc,
@@ -99,7 +98,7 @@ ShardResult<std::map<PropertyId, Value>> FillUpSourceVertexProperties(const std:
     }
     auto pks = PrimaryKeysFromAccessor(*v_acc, view, schema);
     if (pks) {
-      src_vertex_properties.merge(*pks);
+      src_vertex_properties.merge(std::move(*pks));
     }
 
   } else if (req.src_vertex_properties.value().empty()) {
@@ -384,13 +383,10 @@ bool FilterOnEdge(DbAccessor &dba, const storage::v3::VertexAccessor &v_acc, con
 }
 
 ShardResult<msgs::ExpandOneResultRow> GetExpandOneResult(
-    Shard::Accessor &acc, msgs::VertexId src_vertex, const msgs::ExpandOneRequest &req,
+    VertexAccessor v_acc, msgs::VertexId src_vertex, const msgs::ExpandOneRequest &req,
     const EdgeUniquenessFunction &maybe_filter_based_on_edge_uniqueness, const EdgeFiller &edge_filler,
     const Schemas::Schema &schema) {
   /// Fill up source vertex
-  const auto primary_key = ConvertPropertyVector(src_vertex.second);
-  auto v_acc = acc.FindVertex(primary_key, View::NEW);
-
   msgs::Vertex source_vertex = {.id = src_vertex};
   auto maybe_secondary_labels = FillUpSourceVertexSecondaryLabels(v_acc, req);
   if (maybe_secondary_labels.HasError()) {
diff --git a/src/storage/v3/request_helper.hpp b/src/storage/v3/request_helper.hpp
index bbe4894e9..37391cb3d 100644
--- a/src/storage/v3/request_helper.hpp
+++ b/src/storage/v3/request_helper.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -233,7 +233,7 @@ ShardResult<std::map<PropertyId, Value>> CollectAllPropertiesImpl(const TAccesso
                  [](std::pair<const PropertyId, PropertyValue> &pair) {
                    return std::make_pair(pair.first, conversions::FromPropertyValueToValue(std::move(pair.second)));
                  });
-  return ret;
+  return {std::move(ret)};
 }
 }  // namespace impl
 
@@ -247,7 +247,7 @@ EdgeUniquenessFunction InitializeEdgeUniquenessFunction(bool only_unique_neighbo
 EdgeFiller InitializeEdgeFillerFunction(const msgs::ExpandOneRequest &req);
 
 ShardResult<msgs::ExpandOneResultRow> GetExpandOneResult(
-    Shard::Accessor &acc, msgs::VertexId src_vertex, const msgs::ExpandOneRequest &req,
+    VertexAccessor v_acc, msgs::VertexId src_vertex, const msgs::ExpandOneRequest &req,
     const EdgeUniquenessFunction &maybe_filter_based_on_edge_uniqueness, const EdgeFiller &edge_filler,
     const Schemas::Schema &schema);
 
diff --git a/src/storage/v3/shard_manager.hpp b/src/storage/v3/shard_manager.hpp
index 74ec8d2d1..ae7a5f4ef 100644
--- a/src/storage/v3/shard_manager.hpp
+++ b/src/storage/v3/shard_manager.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -250,8 +250,8 @@ class ShardManager {
 
     spdlog::info("SM sending heartbeat to coordinator {}", coordinator_leader_.ToString());
     heartbeat_res_.emplace(std::move(
-        io_.template Request<WriteRequest<CoordinatorWriteRequests>, WriteResponse<CoordinatorWriteResponses>>(
-            coordinator_leader_, ww)));
+        io_.template Request<WriteResponse<CoordinatorWriteResponses>, WriteRequest<CoordinatorWriteRequests>>(
+            coordinator_leader_, std::move(ww))));
     spdlog::info("SM sent heartbeat");
   }
 
diff --git a/src/storage/v3/shard_rsm.cpp b/src/storage/v3/shard_rsm.cpp
index 881796a70..6f71c7269 100644
--- a/src/storage/v3/shard_rsm.cpp
+++ b/src/storage/v3/shard_rsm.cpp
@@ -472,7 +472,8 @@ msgs::ReadResponses ShardRsm::HandleRead(msgs::ExpandOneRequest &&req) {
       if (req.order_by_edges.empty()) {
         const auto *schema = shard_->GetSchema(shard_->PrimaryLabel());
         MG_ASSERT(schema);
-        return GetExpandOneResult(acc, src_vertex, req, maybe_filter_based_on_edge_uniqueness, edge_filler, *schema);
+        return GetExpandOneResult(src_vertex_acc, std::move(src_vertex), req, maybe_filter_based_on_edge_uniqueness,
+                                  edge_filler, *schema);
       }
       auto [in_edge_accessors, out_edge_accessors] = GetEdgesFromVertex(src_vertex_acc, req.direction);
       const auto in_ordered_edges = OrderByEdges(dba, in_edge_accessors, req.order_by_edges, src_vertex_acc);
@@ -487,12 +488,13 @@ msgs::ReadResponses ShardRsm::HandleRead(msgs::ExpandOneRequest &&req) {
                      [](const auto &edge_element) { return edge_element.object_acc; });
       const auto *schema = shard_->GetSchema(shard_->PrimaryLabel());
       MG_ASSERT(schema);
-      return GetExpandOneResult(src_vertex_acc, src_vertex, req, in_edge_ordered_accessors, out_edge_ordered_accessors,
-                                maybe_filter_based_on_edge_uniqueness, edge_filler, *schema);
+      return GetExpandOneResult(src_vertex_acc, std::move(src_vertex), req, std::move(in_edge_ordered_accessors),
+                                std::move(out_edge_ordered_accessors), maybe_filter_based_on_edge_uniqueness,
+                                edge_filler, *schema);
     });
 
     if (maybe_result.HasError()) {
-      shard_error.emplace(CreateErrorResponse(primary_key.GetError(), req.transaction_id, "getting primary key"));
+      shard_error.emplace(CreateErrorResponse(maybe_result.GetError(), req.transaction_id, "getting expand result"));
       break;
     }
 
@@ -581,12 +583,12 @@ msgs::ReadResponses ShardRsm::HandleRead(msgs::GetPropertiesRequest &&req) {
     if (maybe_id.HasError()) {
       return {maybe_id.GetError()};
     }
-    const auto &id = maybe_id.GetValue();
+    auto &vertex_id = maybe_id.GetValue();
     std::optional<msgs::EdgeId> e_id;
     if (e_acc) {
       e_id = msgs::EdgeId{e_acc->Gid().AsUint()};
     }
-    msgs::VertexId v_id{msgs::Label{id.primary_label}, ConvertValueVector(id.primary_key)};
+    msgs::VertexId v_id{msgs::Label{vertex_id.primary_label}, ConvertValueVector(std::move(vertex_id.primary_key))};
     auto maybe_props = collect_props(v_acc, e_acc);
     if (maybe_props.HasError()) {
       return {maybe_props.GetError()};
diff --git a/src/storage/v3/shard_rsm.hpp b/src/storage/v3/shard_rsm.hpp
index d301bf40b..18199d8a1 100644
--- a/src/storage/v3/shard_rsm.hpp
+++ b/src/storage/v3/shard_rsm.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -42,7 +42,7 @@ class ShardRsm {
   explicit ShardRsm(std::unique_ptr<Shard> &&shard) : shard_(std::move(shard)){};
 
   // NOLINTNEXTLINE(readability-convert-member-functions-to-static)
-  msgs::ReadResponses Read(msgs::ReadRequests requests) {
+  msgs::ReadResponses Read(msgs::ReadRequests &&requests) {
     return std::visit([&](auto &&request) mutable { return HandleRead(std::forward<decltype(request)>(request)); },
                       std::move(requests));
   }
diff --git a/src/storage/v3/shard_worker.hpp b/src/storage/v3/shard_worker.hpp
index 547aa0a6f..121c4be70 100644
--- a/src/storage/v3/shard_worker.hpp
+++ b/src/storage/v3/shard_worker.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -100,7 +100,7 @@ class Queue {
 
       inner_->submitted++;
 
-      inner_->queue.emplace_back(std::forward<Message>(message));
+      inner_->queue.emplace_back(std::move(message));
     }  // lock dropped before notifying condition variable
 
     inner_->cv.notify_all();
diff --git a/src/storage/v3/value_conversions.hpp b/src/storage/v3/value_conversions.hpp
index 53374e1ed..c068378ed 100644
--- a/src/storage/v3/value_conversions.hpp
+++ b/src/storage/v3/value_conversions.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -126,6 +126,17 @@ inline std::vector<Value> ConvertValueVector(const std::vector<v3::PropertyValue
   return ret;
 }
 
+inline std::vector<Value> ConvertValueVector(std::vector<v3::PropertyValue> &&vec) {
+  std::vector<Value> ret;
+  ret.reserve(vec.size());
+
+  for (auto &&elem : vec) {
+    ret.push_back(FromPropertyValueToValue(std::move(elem)));
+  }
+
+  return ret;
+}
+
 inline msgs::VertexId ToMsgsVertexId(const v3::VertexId &vertex_id) {
   return {msgs::Label{vertex_id.primary_label}, ConvertValueVector(vertex_id.primary_key)};
 }
diff --git a/src/utils/concepts.hpp b/src/utils/concepts.hpp
index 7c1f3a9c8..9ffc04642 100644
--- a/src/utils/concepts.hpp
+++ b/src/utils/concepts.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -12,6 +12,7 @@
 #pragma once
 #include <concepts>
 #include <iterator>
+#include <type_traits>
 
 namespace memgraph::utils {
 template <typename T, typename... Args>
@@ -34,4 +35,7 @@ template <typename T>
 concept Dereferenceable = requires(T t) {
   { *t } -> CanReference;
 };
+
+template <typename T>
+concept Object = std::is_object_v<T>;
 }  // namespace memgraph::utils
diff --git a/tests/e2e/distributed_queries/awesome_memgraph_functions.py b/tests/e2e/distributed_queries/awesome_memgraph_functions.py
index 0bdaa07a4..0babec2d3 100644
--- a/tests/e2e/distributed_queries/awesome_memgraph_functions.py
+++ b/tests/e2e/distributed_queries/awesome_memgraph_functions.py
@@ -36,7 +36,9 @@ def test_awesome_memgraph_functions(connection):
     assert len(results) == 1
     assert results[0][0] == 5
 
-    results = execute_and_fetch_all(cursor, "MATCH (n) WITH COLLECT(n.property) as nn RETURN ALL(i IN nn WHERE i > 0)")
+    results = execute_and_fetch_all(
+        cursor, "UNWIND [2, 1, 3] AS value WITH COLLECT(value) as nn RETURN ALL(i IN nn WHERE i > 0)"
+    )
     assert len(results) == 1
     assert results[0][0] == True
 
diff --git a/tests/e2e/distributed_queries/distinct.py b/tests/e2e/distributed_queries/distinct.py
index 9ebe50e6f..4a8560b03 100644
--- a/tests/e2e/distributed_queries/distinct.py
+++ b/tests/e2e/distributed_queries/distinct.py
@@ -9,11 +9,13 @@
 # by the Apache License, Version 2.0, included in the file
 # licenses/APL.txt.
 
-import typing
-import mgclient
 import sys
-import pytest
 import time
+import typing
+
+import mgclient
+import pytest
+
 from common import *
 
 
@@ -30,8 +32,7 @@ def test_distinct(connection):
     assert len(results) == 2
     for i, n in enumerate(results):
         n_props = n[0].properties
-        assert len(n_props) == 1
-        assert n_props["property"] == i
+        assert len(n_props) == 0
 
 
 if __name__ == "__main__":
diff --git a/tests/e2e/distributed_queries/distributed_expand_one.py b/tests/e2e/distributed_queries/distributed_expand_one.py
index 9d3d81074..1a74daeb1 100644
--- a/tests/e2e/distributed_queries/distributed_expand_one.py
+++ b/tests/e2e/distributed_queries/distributed_expand_one.py
@@ -13,7 +13,12 @@ import sys
 
 import pytest
 
-from common import connection, execute_and_fetch_all, has_n_result_row, wait_for_shard_manager_to_initialize
+from common import (
+    connection,
+    execute_and_fetch_all,
+    has_n_result_row,
+    wait_for_shard_manager_to_initialize,
+)
 
 
 def test_sequenced_expand_one(connection):
@@ -22,15 +27,21 @@ def test_sequenced_expand_one(connection):
 
     for i in range(1, 4):
         assert has_n_result_row(cursor, f"CREATE (:label {{property:{i}}})", 0), f"Failed creating node"
-    assert has_n_result_row(cursor, "MATCH (n {property:1}), (m {property:2}) CREATE (n)-[:TO]->(m)", 0)
-    assert has_n_result_row(cursor, "MATCH (n {property:2}), (m {property:3}) CREATE (n)-[:TO]->(m)", 0)
+    assert has_n_result_row(cursor, "MATCH (n:label {property:1}), (m:label {property:2}) CREATE (n)-[:TO]->(m)", 0)
+    assert has_n_result_row(cursor, "MATCH (n:label {property:2}), (m:label {property:3}) CREATE (n)-[:TO]->(m)", 0)
 
     results = execute_and_fetch_all(cursor, "MATCH (n)-[:TO]->(m)-[:TO]->(l) RETURN n,m,l")
     assert len(results) == 1
     n, m, l = results[0]
-    assert n.properties["property"] == 1
-    assert m.properties["property"] == 2
-    assert l.properties["property"] == 3
+    assert (
+        len(n.properties) == 0
+    ), "we don't return any properties of the node received from expansion and the bolt layer doesn't serialize the primary key of vertices"
+    assert (
+        len(m.properties) == 0
+    ), "we don't return any properties of the node received from expansion and the bolt layer doesn't serialize the primary key of vertices"
+    assert (
+        len(l.properties) == 0
+    ), "we don't return any properties of the node received from expansion and the bolt layer doesn't serialize the primary key of vertices"
 
 
 if __name__ == "__main__":
diff --git a/tests/e2e/distributed_queries/distributed_queries.py b/tests/e2e/distributed_queries/distributed_queries.py
index b4b6324d9..460aa517c 100644
--- a/tests/e2e/distributed_queries/distributed_queries.py
+++ b/tests/e2e/distributed_queries/distributed_queries.py
@@ -9,11 +9,13 @@
 # by the Apache License, Version 2.0, included in the file
 # licenses/APL.txt.
 
-import typing
-import mgclient
 import sys
-import pytest
 import time
+import typing
+
+import mgclient
+import pytest
+
 from common import *
 
 
@@ -35,13 +37,13 @@ def test_vertex_creation_and_scanall(connection):
     assert len(results) == 9
     for (n, r, m) in results:
         n_props = n.properties
-        assert len(n_props) == 1, "n is not expected to have properties, update the test!"
+        assert len(n_props) == 0, "n is not expected to have properties, update the test!"
         assert len(n.labels) == 0, "n is not expected to have labels, update the test!"
 
         assert r.type == "TO"
 
         m_props = m.properties
-        assert m_props["property"] <= 3 and m_props["property"] >= 0, "Wrong key"
+        assert len(m_props) == 0, "n is not expected to have properties, update the test!"
         assert len(m.labels) == 0, "m is not expected to have labels, update the test!"
 
 
diff --git a/tests/e2e/distributed_queries/order_by_and_limit.py b/tests/e2e/distributed_queries/order_by_and_limit.py
index 05297f8f6..29be77857 100644
--- a/tests/e2e/distributed_queries/order_by_and_limit.py
+++ b/tests/e2e/distributed_queries/order_by_and_limit.py
@@ -9,11 +9,13 @@
 # by the Apache License, Version 2.0, included in the file
 # licenses/APL.txt.
 
-import typing
-import mgclient
 import sys
-import pytest
 import time
+import typing
+
+import mgclient
+import pytest
+
 from common import *
 
 
@@ -21,23 +23,23 @@ def test_order_by_and_limit(connection):
     wait_for_shard_manager_to_initialize()
     cursor = connection.cursor()
 
-    assert has_n_result_row(cursor, "CREATE (n :label {property:1})", 0)
-    assert has_n_result_row(cursor, "CREATE (n :label {property:2})", 0)
-    assert has_n_result_row(cursor, "CREATE (n :label {property:3})", 0)
-    assert has_n_result_row(cursor, "CREATE (n :label {property:4})", 0)
-
-    results = execute_and_fetch_all(cursor, "MATCH (n) RETURN n ORDER BY n.property DESC")
-    assert len(results) == 4
-    i = 4
-    for n in results:
-        n_props = n[0].properties
-        assert len(n_props) == 1
-        assert n_props["property"] == i
+    results = execute_and_fetch_all(
+        cursor,
+        "UNWIND [{property:1}, {property:3}, {property:2}] AS map RETURN map ORDER BY map.property DESC",
+    )
+    assert len(results) == 3
+    i = 3
+    for map in results:
+        assert len(map) == 1
+        assert map[0]["property"] == i
         i = i - 1
 
-    result = execute_and_fetch_all(cursor, "MATCH (n) RETURN n ORDER BY n.property LIMIT 1")
+    result = execute_and_fetch_all(
+        cursor,
+        "UNWIND [{property:1}, {property:3}, {property:2}] AS map RETURN map ORDER BY map.property LIMIT 1",
+    )
     assert len(result) == 1
-    assert result[0][0].properties["property"] == 1
+    assert result[0][0]["property"] == 1
 
 
 if __name__ == "__main__":
diff --git a/tests/manual/CMakeLists.txt b/tests/manual/CMakeLists.txt
index ea680105c..800cd0515 100644
--- a/tests/manual/CMakeLists.txt
+++ b/tests/manual/CMakeLists.txt
@@ -9,6 +9,7 @@ function(add_manual_test test_cpp)
   get_filename_component(exec_name ${test_cpp} NAME_WE)
   set(target_name ${test_prefix}${exec_name})
   add_executable(${target_name} ${test_cpp} ${ARGN})
+
   # OUTPUT_NAME sets the real name of a target when it is built and can be
   # used to help create two targets of the same name even though CMake
   # requires unique logical target names
@@ -21,7 +22,7 @@ target_link_libraries(${test_prefix}antlr_parser antlr_opencypher_parser_lib)
 
 add_manual_test(antlr_sigsegv.cpp)
 target_link_libraries(${test_prefix}antlr_sigsegv gtest gtest_main
-                      antlr_opencypher_parser_lib mg-utils)
+  antlr_opencypher_parser_lib mg-utils)
 
 add_manual_test(antlr_tree_pretty_print.cpp)
 target_link_libraries(${test_prefix}antlr_tree_pretty_print antlr_opencypher_parser_lib)
@@ -37,13 +38,15 @@ target_link_libraries(${test_prefix}query_hash mg-query)
 
 add_manual_test(query_planner.cpp interactive/planning.cpp)
 target_link_libraries(${test_prefix}query_planner mg-query)
-if (READLINE_FOUND)
+
+if(READLINE_FOUND)
   target_link_libraries(${test_prefix}query_planner readline)
 endif()
 
 add_manual_test(query_execution_dummy.cpp)
 target_link_libraries(${test_prefix}query_execution_dummy mg-query)
-if (READLINE_FOUND)
+
+if(READLINE_FOUND)
   target_link_libraries(${test_prefix}query_execution_dummy readline)
 endif()
 
@@ -61,3 +64,6 @@ target_link_libraries(${test_prefix}ssl_client mg-communication)
 
 add_manual_test(ssl_server.cpp)
 target_link_libraries(${test_prefix}ssl_server mg-communication)
+
+add_manual_test(query_performance.cpp)
+target_link_libraries(${test_prefix}query_performance mg-communication mg-utils mg-io mg-io-simulator mg-coordinator mg-query-v2 mg-storage-v3 mg-query mg-storage-v2)
diff --git a/tests/manual/query_performance.cpp b/tests/manual/query_performance.cpp
new file mode 100644
index 000000000..8c7877b0d
--- /dev/null
+++ b/tests/manual/query_performance.cpp
@@ -0,0 +1,352 @@
+// Copyright 2023 Memgraph Ltd.
+//
+// Use of this software is governed by the Business Source License
+// included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
+// License, and you may not use this file except in compliance with the Business Source License.
+//
+// As of the Change Date specified in that file, in accordance with
+// the Business Source License, use of this software will be governed
+// by the Apache License, Version 2.0, included in the file
+// licenses/APL.txt.
+
+// This binary is meant to easily compare the performance of:
+//  - Memgraph v2
+//  - Memgraph v3
+//  - Memgraph v3 with MultiFrame
+// This binary measures three things which provides a high level and easily understandable metric about the performance
+// difference between the different versions:
+//  1. Read time: how much time does it take to read the files:
+//  2. Init time: how much time does it take to run the init queries, including the index creation. For details please
+//     check RunV2.
+//  3. Benchmark time: how much time does it take to run the benchmark queries.
+// To quickly compare performance of the different versions just change the query or queries in the benchmark queries
+// file you can see the different by running this executable. This way we don't have keep multiple binaries of Memgraph
+// v2 and Memgraph v3 with/without MultiFrame, start Memgraph and connect to it with mgconsole and other hassles. As
+// everything is run in this binary, it makes easier to generate perf reports/flamegraphs from the query execution of
+// different Memgraph versions compared to using the full blown version of Memgraph.
+//
+// A few important notes:
+//  - All the input files are mandated to have an empty line at the end of the file as the reading logic expect that.
+//  - tests/mgbench/dataset_creator_unwind.py is recommended to generate the dataset because it generates queries with
+//    UNWIND that makes the import faster in Memgraph v3, thus we can compare the performance on non trivial datasets
+//    also. To make it possible to use the generated dataset, you have to move the generated index queries into a
+//    separate file that can be supplied as index queries file for this binary when using Memgraph v2. The reason for
+//    this is Memgraph v3 cannot handle indices yet, thus it crashes.
+//  - Check the command line flags and their description defined in this file.
+//  - Also check out the --default-multi-frame-size command line flag if you want to play with that.
+//  - The log level is manually set to warning in the main function to avoid the overwhelming log messages from Memgraph
+//    v3. Apart from ease of use, the huge amount of looging can degrade the actual performance.
+//
+// Example usage with Memgraph v2:
+// ./query_performance
+//   --index-queries-file indices.cypher
+//   --init-queries-file dataset.cypher
+//   --benchmark-queries-files expand.cypher,match.cypyher
+//   --use-v3=false
+//
+// Example usage with Memgraph v3 without MultiFrame:
+// ./query_performance
+//   --split-file split_file
+//   --init-queries-file dataset.cypher
+//   --benchmark-queries-files expand.cypher,match.cypyher
+//   --use-v3=true
+//   --use-multi-frame=false
+//
+// Example usage with Memgraph v3 with MultiFrame:
+// ./query_performance
+//   --split-file split_file
+//   --init-queries-file dataset.cypher
+//   --benchmark-queries-files expand.cypher,match.cypyher
+//   --use-v3=true
+//   --use-multi-frame=true
+//
+// The examples are using only the necessary flags, however specifying all of them is not a problem, so if you specify
+// --index-queries-file for Memgraph v3, then it will be safely ignored just as --split-file for Memgraph v2.
+//
+// To generate flamegraph you can use the following command:
+// flamegraph --cmd "record -F 997 --call-graph fp -g" --root -o flamegraph.svg -- ./query_performance <flags>
+// Using the default option (dwarf) for --call-graph when calling perf might result in too long runtine of flamegraph
+// because of address resolution. See https://github.com/flamegraph-rs/flamegraph/issues/74.
+
+#include <chrono>
+#include <filesystem>
+#include <fstream>
+#include <istream>
+#include <thread>
+
+#include <fmt/core.h>
+#include <gflags/gflags.h>
+#include <spdlog/cfg/env.h>
+#include <spdlog/spdlog.h>
+#include <json/json.hpp>
+
+// v3 includes
+#include "io/address.hpp"
+#include "io/local_transport/local_system.hpp"
+#include "io/message_histogram_collector.hpp"
+#include "machine_manager/machine_manager.hpp"
+#include "query/discard_value_stream.hpp"
+#include "query/v2/discard_value_stream.hpp"
+#include "query/v2/interpreter.hpp"
+#include "query/v2/request_router.hpp"
+
+// v2 includes
+#include "query/interpreter.hpp"
+#include "storage/v2/storage.hpp"
+
+// common includes
+#include "utils/string.hpp"
+
+// NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)
+DEFINE_string(index_queries_file, "",
+              "Path to the file which contains the queries to create indices. Used only for v2. Must contain an empty "
+              "line at the end of the file after the queries.");
+// NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)
+DEFINE_string(split_file, "",
+              "Path to the split file which contains the predefined labels, properties, edge types and shard-ranges. "
+              "Used only for v3. Must contain an empty line at the end of the file.");
+// NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)
+DEFINE_string(init_queries_file, "",
+              "Path to the file that is used to insert the initial dataset, one query per line. Must contain an empty "
+              "line at the end of the file after the queries.");
+// NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)
+DEFINE_string(benchmark_queries_files, "",
+              "Comma separated paths to the files that contain the queries that we want to compare, one query per "
+              "line. Must contain an empty line at the end of each file after the queries.");
+// NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)
+DEFINE_bool(use_v3, true, "If set to true, then Memgraph v3 will be used, otherwise Memgraph v2 will be used.");
+
+DEFINE_string(export_json_results, "", "If not empty, then the results will be exported as a json file.");
+
+DEFINE_string(data_directory, "mg_data", "Path to directory to use as storage directory for Memgraph v2.");
+
+namespace memgraph::tests::manual {
+
+template <typename TInterpreterContext>
+struct DependantTypes {};
+
+template <>
+struct DependantTypes<query::InterpreterContext> {
+  using Interpreter = query::Interpreter;
+  using DiscardValueResultStream = query::DiscardValueResultStream;
+};
+
+template <>
+struct DependantTypes<query::v2::InterpreterContext> {
+  using Interpreter = query::v2::Interpreter;
+  using DiscardValueResultStream = query::v2::DiscardValueResultStream;
+};
+
+template <typename TRep, typename TPeriod>
+void PutResult(nlohmann::json &json, const std::string_view name, std::chrono::duration<TRep, TPeriod> duration) {
+  json[name] = std::chrono::duration_cast<std::chrono::microseconds>(duration).count();
+}
+
+template <typename TInterpreterContext>
+using Interpreter = typename DependantTypes<TInterpreterContext>::Interpreter;
+
+template <typename TInterpreterContext>
+using DiscardValueResultStream = typename DependantTypes<TInterpreterContext>::DiscardValueResultStream;
+
+template <typename TInterpreterContext>
+void RunQueries(TInterpreterContext &interpreter_context, const std::vector<std::string> &queries) {
+  Interpreter<TInterpreterContext> interpreter{&interpreter_context};
+  DiscardValueResultStream<TInterpreterContext> stream;
+
+  for (const auto &query : queries) {
+    auto result = interpreter.Prepare(query, {}, nullptr);
+    interpreter.Pull(&stream, std::nullopt, result.qid);
+  }
+}
+
+template <typename TInterpreterContext>
+void RunInitQueries(TInterpreterContext &interpreter_context, const std::vector<std::string> &init_queries) {
+  RunQueries(interpreter_context, init_queries);
+}
+
+template <typename TInterpreterContext>
+void RunBenchmarkQueries(TInterpreterContext &interpreter_context, const std::vector<std::string> &benchmark_queries) {
+  RunQueries(interpreter_context, benchmark_queries);
+}
+
+std::vector<std::string> ReadQueries(const std::string &file_name) {
+  std::vector<std::string> queries{};
+  std::string buffer;
+
+  std::ifstream file{file_name, std::ios::in};
+  MG_ASSERT(file.good(), "Cannot open queries file to read: {}", file_name);
+  while (file.good()) {
+    std::getline(file, buffer);
+    if (buffer.empty()) {
+      continue;
+    }
+    // Trim the trailing `;`
+    queries.push_back(buffer.substr(0, buffer.size() - 1));
+  }
+  return queries;
+}
+
+std::map<std::string, std::vector<std::string>> ReadBenchmarkQueries(const std::string benchmark_queries_files) {
+  auto benchmark_files = utils::Split(benchmark_queries_files, ",");
+  std::map<std::string, std::vector<std::string>> result;
+  for (const auto &benchmark_file : benchmark_files) {
+    const auto path = std::filesystem::path(benchmark_file);
+    result.emplace(path.stem().string(), ReadQueries(benchmark_file));
+  }
+  return result;
+}
+
+void RunV2() {
+  spdlog::critical("Running V2");
+  const auto run_start = std::chrono::high_resolution_clock::now();
+
+  const auto index_queries = ReadQueries(FLAGS_index_queries_file);
+  const auto init_queries = ReadQueries(FLAGS_init_queries_file);
+  const auto benchmarks = ReadBenchmarkQueries(FLAGS_benchmark_queries_files);
+
+  storage::Storage storage{
+      storage::Config{.durability{.storage_directory = FLAGS_data_directory,
+                                  .snapshot_wal_mode = storage::Config::Durability::SnapshotWalMode::DISABLED}}};
+
+  memgraph::query::InterpreterContext interpreter_context{
+      &storage,
+      {.query = {.allow_load_csv = false},
+       .execution_timeout_sec = 0,
+       .replication_replica_check_frequency = std::chrono::seconds(0),
+       .default_kafka_bootstrap_servers = "",
+       .default_pulsar_service_url = "",
+       .stream_transaction_conflict_retries = 0,
+       .stream_transaction_retry_interval = std::chrono::milliseconds(0)},
+      FLAGS_data_directory};
+
+  const auto init_start = std::chrono::high_resolution_clock::now();
+  RunInitQueries(interpreter_context, index_queries);
+  RunInitQueries(interpreter_context, init_queries);
+  const auto benchmark_start = std::chrono::high_resolution_clock::now();
+
+  spdlog::critical("Read: {}ms", std::chrono::duration_cast<std::chrono::milliseconds>(init_start - run_start).count());
+  spdlog::critical("Init: {}ms",
+                   std::chrono::duration_cast<std::chrono::milliseconds>(benchmark_start - init_start).count());
+
+  std::map<std::string, std::chrono::nanoseconds> benchmark_results;
+  for (const auto &[name, queries] : benchmarks) {
+    const auto current_start = std::chrono::high_resolution_clock::now();
+    RunBenchmarkQueries(interpreter_context, queries);
+    const auto current_stop = std::chrono::high_resolution_clock::now();
+    const auto elapsed = current_stop - current_start;
+    spdlog::critical("Benchmark {}: {}ms", name,
+                     std::chrono::duration_cast<std::chrono::milliseconds>(elapsed).count());
+    benchmark_results.emplace(name, elapsed);
+  }
+
+  const auto benchmark_end = std::chrono::high_resolution_clock::now();
+  spdlog::critical("Benchmark: {}ms",
+                   std::chrono::duration_cast<std::chrono::milliseconds>(benchmark_end - benchmark_start).count());
+
+  if (!FLAGS_export_json_results.empty()) {
+    nlohmann::json results;
+    PutResult(results, "init", benchmark_start - init_start);
+    nlohmann::json benchmark_results_json;
+    for (const auto &[name, duration] : benchmark_results) {
+      PutResult(benchmark_results_json, name, duration);
+    }
+    results["benchmarks"] = std::move(benchmark_results_json);
+    std::ofstream results_file{FLAGS_export_json_results};
+    results_file << results.dump();
+  }
+}
+
+void RunV3() {
+  spdlog::critical("Running V3");
+  const auto run_start = std::chrono::high_resolution_clock::now();
+  std::ifstream sm_file{FLAGS_split_file, std::ios::in};
+  MG_ASSERT(sm_file.good(), "Cannot open split file to read: {}", FLAGS_split_file);
+  auto sm = memgraph::coordinator::ShardMap::Parse(sm_file);
+
+  const auto init_queries = ReadQueries(FLAGS_init_queries_file);
+  const auto benchmarks = ReadBenchmarkQueries(FLAGS_benchmark_queries_files);
+
+  io::local_transport::LocalSystem ls;
+
+  auto unique_local_addr_query = io::Address::UniqueLocalAddress();
+  auto io = ls.Register(unique_local_addr_query);
+
+  memgraph::machine_manager::MachineConfig config{
+      .coordinator_addresses = std::vector<memgraph::io::Address>{unique_local_addr_query},
+      .is_storage = true,
+      .is_coordinator = true,
+      .listen_ip = unique_local_addr_query.last_known_ip,
+      .listen_port = unique_local_addr_query.last_known_port,
+      .shard_worker_threads = 2,
+  };
+
+  memgraph::coordinator::Coordinator coordinator{sm};
+
+  memgraph::machine_manager::MachineManager<memgraph::io::local_transport::LocalTransport> mm{io, config, coordinator};
+  std::jthread mm_thread([&mm] { mm.Run(); });
+
+  auto rr_factory = std::make_unique<memgraph::query::v2::LocalRequestRouterFactory>(io);
+
+  query::v2::InterpreterContext interpreter_context{(memgraph::storage::v3::Shard *)(nullptr),
+                                                    {.execution_timeout_sec = 0},
+                                                    "data",
+                                                    std::move(rr_factory),
+                                                    mm.CoordinatorAddress()};
+
+  // without this it fails sometimes because the CreateVertices request might reach the shard worker faster than the
+  // ShardToInitialize
+  std::this_thread::sleep_for(std::chrono::milliseconds(150));
+
+  const auto init_start = std::chrono::high_resolution_clock::now();
+  RunInitQueries(interpreter_context, init_queries);
+  const auto benchmark_start = std::chrono::high_resolution_clock::now();
+
+  spdlog::critical("Read: {}ms", std::chrono::duration_cast<std::chrono::milliseconds>(init_start - run_start).count());
+  spdlog::critical("Init: {}ms",
+                   std::chrono::duration_cast<std::chrono::milliseconds>(benchmark_start - init_start).count());
+
+  std::map<std::string, std::chrono::nanoseconds> benchmark_results;
+  for (const auto &[name, queries] : benchmarks) {
+    const auto current_start = std::chrono::high_resolution_clock::now();
+    RunBenchmarkQueries(interpreter_context, queries);
+    const auto current_stop = std::chrono::high_resolution_clock::now();
+    const auto elapsed = current_stop - current_start;
+    spdlog::critical("Benchmark {}: {}ms", name,
+                     std::chrono::duration_cast<std::chrono::milliseconds>(elapsed).count());
+    benchmark_results.emplace(name, elapsed);
+  }
+
+  const auto benchmark_end = std::chrono::high_resolution_clock::now();
+  spdlog::critical("Benchmark: {}ms",
+                   std::chrono::duration_cast<std::chrono::milliseconds>(benchmark_end - benchmark_start).count());
+
+  ls.ShutDown();
+  auto latency_histograms = nlohmann::json::parse(fmt::format("{}", io.ResponseLatencies()));
+  spdlog::warn(latency_histograms.dump(4));
+
+  if (!FLAGS_export_json_results.empty()) {
+    nlohmann::json results;
+    PutResult(results, "init", benchmark_start - init_start);
+    nlohmann::json benchmark_results_json;
+    for (const auto &[name, duration] : benchmark_results) {
+      PutResult(benchmark_results_json, name, duration);
+    }
+    results["benchmarks"] = std::move(benchmark_results_json);
+    results["latencies"] = std::move(latency_histograms);
+    std::ofstream results_file{FLAGS_export_json_results};
+    results_file << results.dump();
+  }
+}
+}  // namespace memgraph::tests::manual
+
+int main(int argc, char **argv) {
+  spdlog::set_level(spdlog::level::warn);
+  spdlog::cfg::load_env_levels();
+  gflags::ParseCommandLineFlags(&argc, &argv, true);
+  if (FLAGS_use_v3) {
+    memgraph::tests::manual::RunV3();
+  } else {
+    memgraph::tests::manual::RunV2();
+  }
+  return 0;
+}
diff --git a/tests/manual/query_performance_runner.py b/tests/manual/query_performance_runner.py
new file mode 100755
index 000000000..fa8b1b4b3
--- /dev/null
+++ b/tests/manual/query_performance_runner.py
@@ -0,0 +1,116 @@
+#!/usr/bin/env python3
+
+# Copyright 2023 Memgraph Ltd.
+#
+# Use of this software is governed by the Business Source License
+# included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
+# License, and you may not use this file except in compliance with the Business Source License.
+#
+# As of the Change Date specified in that file, in accordance with
+# the Business Source License, use of this software will be governed
+# by the Apache License, Version 2.0, included in the file
+# licenses/APL.txt.
+
+import argparse
+import io
+import json
+import os
+import subprocess
+import tarfile
+import tempfile
+
+import requests
+
+SCRIPT_DIR = os.path.dirname(os.path.realpath(__file__))
+PROJECT_DIR = os.path.normpath(os.path.join(SCRIPT_DIR, "..", ".."))
+BUILD_DIR = os.path.join(PROJECT_DIR, "build")
+BINARY_DIR = os.path.join(BUILD_DIR, "tests/manual")
+DEFAULT_BENCHMARK_DIR = os.path.join(BINARY_DIR, "query_performance_benchmark")
+DATA_URL = (
+    "https://s3.eu-west-1.amazonaws.com/deps.memgraph.io/dataset/query_performance/query_performance_benchmark.tar.gz"
+)
+
+parser = argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+parser.add_argument(
+    "--binary",
+    type=str,
+    default=os.path.join(BINARY_DIR, "query_performance"),
+    help="Path to the binary to use for the benchmark.",
+)
+parser.add_argument(
+    "--data-dir",
+    type=str,
+    default=tempfile.TemporaryDirectory().name,
+    help="Path to directory that can be used as a data directory for ",
+)
+parser.add_argument(
+    "--summary-path",
+    type=str,
+    default=os.path.join(DEFAULT_BENCHMARK_DIR, "summary.json"),
+    help="Path to which file write the summary.",
+)
+
+parser.add_argument("--init-queries-file", type=str, default=os.path.join(DEFAULT_BENCHMARK_DIR, "dataset.cypher"))
+parser.add_argument("--index-queries-file", type=str, default=os.path.join(DEFAULT_BENCHMARK_DIR, "indices.cypher"))
+parser.add_argument("--split-file", type=str, default=os.path.join(DEFAULT_BENCHMARK_DIR, "split_file"))
+
+parser.add_argument(
+    "--benchmark-queries-files",
+    type=str,
+    default=",".join(
+        [os.path.join(DEFAULT_BENCHMARK_DIR, file_name) for file_name in ["expand.cypher", "match_files.cypher"]]
+    ),
+)
+
+args = parser.parse_args()
+
+v2_results_path = os.path.join(DEFAULT_BENCHMARK_DIR, "v2_results.json")
+v3_results_path = os.path.join(DEFAULT_BENCHMARK_DIR, "v3_results.json")
+
+
+if os.path.exists(DEFAULT_BENCHMARK_DIR):
+    print(f"Using cachced data from {DEFAULT_BENCHMARK_DIR}")
+else:
+    print(f"Downloading benchmark data to {DEFAULT_BENCHMARK_DIR}")
+    r = requests.get(DATA_URL)
+    assert r.ok, "Cannot download data"
+    file_like_object = io.BytesIO(r.content)
+    tar = tarfile.open(fileobj=file_like_object)
+    tar.extractall(os.path.dirname(DEFAULT_BENCHMARK_DIR))
+
+subprocess.run(
+    [
+        args.binary,
+        f"--split-file={args.split_file}",
+        f"--index-queries-file={args.index_queries_file}",
+        f"--init-queries-file={args.init_queries_file}",
+        f"--benchmark-queries-files={args.benchmark_queries_files}",
+        "--use-v3=false",
+        "--use-multi-frame=true",
+        f"--export-json-results={v2_results_path}",
+        f"--data-directory={args.data_dir}",
+    ]
+)
+
+subprocess.run(
+    [
+        args.binary,
+        f"--split-file={args.split_file}",
+        f"--index-queries-file={args.index_queries_file}",
+        f"--init-queries-file={args.init_queries_file}",
+        f"--benchmark-queries-files={args.benchmark_queries_files}",
+        "--use-v3=true",
+        "--use-multi-frame=true",
+        f"--export-json-results={v3_results_path}",
+        f"--data-directory={args.data_dir}",
+    ]
+)
+
+
+v2_results_file = open(v2_results_path)
+v2_results = json.load(v2_results_file)
+v3_results_file = open(v3_results_path)
+v3_results = json.load(v3_results_file)
+
+with open(args.summary_path, "w") as summary:
+    json.dump({"v2": v2_results, "v3": v3_results}, summary)
diff --git a/tests/mgbench/dataset_creator.py b/tests/mgbench/dataset_creator.py
index 9ebeb8cd1..0b73eceed 100644
--- a/tests/mgbench/dataset_creator.py
+++ b/tests/mgbench/dataset_creator.py
@@ -51,10 +51,22 @@ import helpers
 
 def main():
     parser = argparse.ArgumentParser()
-    parser.add_argument("--number_of_identities", type=int, default=10)
-    parser.add_argument("--number_of_files", type=int, default=10)
-    parser.add_argument("--percentage_of_permissions", type=float, default=1.0)
-    parser.add_argument("--filename", default="dataset.cypher")
+    parser.add_argument(
+        "--number_of_identities",
+        type=int,
+        default=10,
+        help="Determines how many :Identity nodes will the dataset contain.",
+    )
+    parser.add_argument(
+        "--number_of_files", type=int, default=10, help="Determines how many :File nodes will the dataset contain."
+    )
+    parser.add_argument(
+        "--percentage_of_permissions",
+        type=float,
+        default=1.0,
+        help="Determines approximately what percentage of the all possible identity-permission-file connections will be created.",
+    )
+    parser.add_argument("--filename", default="dataset.cypher", help="The name of the output file.")
 
     args = parser.parse_args()
 
diff --git a/tests/mgbench/dataset_creator_unwind.py b/tests/mgbench/dataset_creator_unwind.py
index 564a4d018..00de1c7bf 100644
--- a/tests/mgbench/dataset_creator_unwind.py
+++ b/tests/mgbench/dataset_creator_unwind.py
@@ -51,10 +51,22 @@ import helpers
 
 def main():
     parser = argparse.ArgumentParser()
-    parser.add_argument("--number_of_identities", type=int, default=10)
-    parser.add_argument("--number_of_files", type=int, default=10)
-    parser.add_argument("--percentage_of_permissions", type=float, default=1.0)
-    parser.add_argument("--filename", default="dataset.cypher")
+    parser.add_argument(
+        "--number_of_identities",
+        type=int,
+        default=10,
+        help="Determines how many :Identity nodes will the dataset contain.",
+    )
+    parser.add_argument(
+        "--number_of_files", type=int, default=10, help="Determines how many :File nodes will the dataset contain."
+    )
+    parser.add_argument(
+        "--percentage_of_permissions",
+        type=float,
+        default=1.0,
+        help="Determines approximately what percentage of the all possible identity-permission-file connections will be created.",
+    )
+    parser.add_argument("--filename", default="dataset.cypher", help="The name of the output file.")
 
     args = parser.parse_args()
 
diff --git a/tests/simulation/basic_request.cpp b/tests/simulation/basic_request.cpp
index 868f4ac10..d6504a365 100644
--- a/tests/simulation/basic_request.cpp
+++ b/tests/simulation/basic_request.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -55,7 +55,7 @@ void run_server(Io<SimulatorTransport> io) {
     highest_seen = std::max(highest_seen, req.proposal);
     auto srv_res = CounterResponse{highest_seen};
 
-    io.Send(request_envelope.from_address, request_envelope.request_id, srv_res);
+    io.Send(request_envelope.from_address, request_envelope.request_id, std::move(srv_res));
   }
 }
 
@@ -76,7 +76,7 @@ std::pair<SimulatorStats, LatencyHistogramSummaries> RunWorkload(SimulatorConfig
     CounterRequest cli_req;
     cli_req.proposal = i;
     spdlog::info("[CLIENT] calling Request");
-    auto res_f = cli_io.Request<CounterRequest, CounterResponse>(srv_addr, cli_req);
+    auto res_f = cli_io.Request<CounterResponse, CounterRequest>(srv_addr, std::move(cli_req));
     spdlog::info("[CLIENT] calling Wait");
     auto res_rez = std::move(res_f).Wait();
     spdlog::info("[CLIENT] Wait returned");
diff --git a/tests/simulation/trial_query_storage/query_storage_test.cpp b/tests/simulation/trial_query_storage/query_storage_test.cpp
index 8ef12bdb8..9e778f59d 100644
--- a/tests/simulation/trial_query_storage/query_storage_test.cpp
+++ b/tests/simulation/trial_query_storage/query_storage_test.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -44,7 +44,7 @@ void run_server(Io<SimulatorTransport> io) {
     for (auto index = start_index; index < start_index + req.count; ++index) {
       response.vertices.push_back({std::string("Vertex_") + std::to_string(index)});
     }
-    io.Send(request_envelope.from_address, request_envelope.request_id, response);
+    io.Send(request_envelope.from_address, request_envelope.request_id, std::move(response));
   }
 }
 
@@ -78,7 +78,7 @@ int main() {
 
   auto req = ScanVerticesRequest{2, std::nullopt};
 
-  auto res_f = cli_io.Request<ScanVerticesRequest, VerticesResponse>(srv_addr, req);
+  auto res_f = cli_io.Request<VerticesResponse, ScanVerticesRequest>(srv_addr, std::move(req));
   auto res_rez = std::move(res_f).Wait();
   simulator.ShutDown();
   return 0;
diff --git a/tests/unit/local_transport.cpp b/tests/unit/local_transport.cpp
index aa03325de..f2d3f8458 100644
--- a/tests/unit/local_transport.cpp
+++ b/tests/unit/local_transport.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
@@ -48,7 +48,7 @@ void RunServer(Io<LocalTransport> io) {
     highest_seen = std::max(highest_seen, req.proposal);
     auto srv_res = CounterResponse{highest_seen};
 
-    io.Send(request_envelope.from_address, request_envelope.request_id, srv_res);
+    io.Send(request_envelope.from_address, request_envelope.request_id, std::move(srv_res));
   }
 }
 
@@ -70,7 +70,7 @@ TEST(LocalTransport, BasicRequest) {
     auto value = 1;  // i;
     cli_req.proposal = value;
     spdlog::info("[CLIENT] sending request");
-    auto res_f = cli_io.Request<CounterRequest, CounterResponse>(srv_addr, cli_req);
+    auto res_f = cli_io.Request<CounterResponse, CounterRequest>(srv_addr, std::move(cli_req));
     spdlog::info("[CLIENT] waiting on future");
 
     auto res_rez = std::move(res_f).Wait();