diff --git a/nomos-mix/core/src/message_blend/crypto.rs b/nomos-mix/core/src/message_blend/crypto.rs
index 62a8a865..b214a39c 100644
--- a/nomos-mix/core/src/message_blend/crypto.rs
+++ b/nomos-mix/core/src/message_blend/crypto.rs
@@ -12,6 +12,7 @@ where
     settings: CryptographicProcessorSettings<M::PrivateKey>,
     membership: Membership<M>,
     rng: R,
+    mix_message: M,
 }
 
 #[derive(Clone, Debug, Serialize, Deserialize)]
@@ -28,18 +29,20 @@ where
 {
     pub fn new(
         settings: CryptographicProcessorSettings<M::PrivateKey>,
+        mix_message_settings: M::Settings,
         membership: Membership<M>,
         rng: R,
     ) -> Self {
+        let mix_message = M::new(mix_message_settings);
         Self {
             settings,
             membership,
             rng,
+            mix_message,
         }
     }
 
     pub fn wrap_message(&mut self, message: &[u8]) -> Result<Vec<u8>, M::Error> {
-        // TODO: Use the actual Sphinx encoding instead of mock.
         let public_keys = self
             .membership
             .choose_remote_nodes(&mut self.rng, self.settings.num_mix_layers)
@@ -47,10 +50,11 @@ where
             .map(|node| node.public_key.clone())
             .collect::<Vec<_>>();
 
-        M::build_message(message, &public_keys)
+        self.mix_message.build_message(message, &public_keys)
     }
 
     pub fn unwrap_message(&self, message: &[u8]) -> Result<(Vec<u8>, bool), M::Error> {
-        M::unwrap_message(message, &self.settings.private_key)
+        self.mix_message
+            .unwrap_message(message, &self.settings.private_key)
     }
 }
diff --git a/nomos-mix/core/src/message_blend/mod.rs b/nomos-mix/core/src/message_blend/mod.rs
index 2cc988b6..140f23ae 100644
--- a/nomos-mix/core/src/message_blend/mod.rs
+++ b/nomos-mix/core/src/message_blend/mod.rs
@@ -59,12 +59,14 @@ where
     pub fn new(
         input_stream: S,
         settings: MessageBlendSettings<M>,
+        mix_message_settings: M::Settings,
         membership: Membership<M>,
         scheduler: Scheduler,
         cryptographic_processor_rng: Rng,
     ) -> Self {
         let cryptographic_processor = CryptographicProcessor::new(
             settings.cryptographic_processor,
+            mix_message_settings,
             membership,
             cryptographic_processor_rng,
         );
@@ -132,6 +134,7 @@ where
     fn blend(
         self,
         message_blend_settings: MessageBlendSettings<M>,
+        mix_message_settings: M::Settings,
         membership: Membership<M>,
         scheduler: Scheduler,
         cryptographic_processor_rng: Rng,
@@ -142,6 +145,7 @@ where
         MessageBlendStream::new(
             self,
             message_blend_settings,
+            mix_message_settings,
             membership,
             scheduler,
             cryptographic_processor_rng,
diff --git a/nomos-mix/core/src/persistent_transmission.rs b/nomos-mix/core/src/persistent_transmission.rs
index 72531a49..1e8e3292 100644
--- a/nomos-mix/core/src/persistent_transmission.rs
+++ b/nomos-mix/core/src/persistent_transmission.rs
@@ -1,9 +1,6 @@
 use futures::{Stream, StreamExt};
-use nomos_mix_message::MixMessage;
 use rand::{distributions::Uniform, prelude::Distribution, Rng, RngCore};
-use serde::de::DeserializeOwned;
 use serde::{Deserialize, Serialize};
-use std::marker::PhantomData;
 use std::pin::{pin, Pin};
 use std::task::{Context, Poll};
 
@@ -25,7 +22,7 @@ impl Default for PersistentTransmissionSettings {
 }
 
 /// Transmit scheduled messages with a persistent rate as a stream.
-pub struct PersistentTransmissionStream<S, Rng, M, Scheduler>
+pub struct PersistentTransmissionStream<S, Rng, Scheduler>
 where
     S: Stream,
     Rng: RngCore,
@@ -33,14 +30,13 @@ where
     coin: Coin<Rng>,
     stream: S,
     scheduler: Scheduler,
-    _mix_message: PhantomData<M>,
+    drop_message: Vec<u8>,
 }
 
-impl<S, Rng, M, Scheduler> PersistentTransmissionStream<S, Rng, M, Scheduler>
+impl<S, Rng, Scheduler> PersistentTransmissionStream<S, Rng, Scheduler>
 where
     S: Stream,
     Rng: RngCore,
-    M: MixMessage,
     Scheduler: Stream<Item = ()>,
 {
     pub fn new(
@@ -48,22 +44,22 @@ where
         stream: S,
         scheduler: Scheduler,
         rng: Rng,
-    ) -> PersistentTransmissionStream<S, Rng, M, Scheduler> {
+        drop_message: Vec<u8>,
+    ) -> PersistentTransmissionStream<S, Rng, Scheduler> {
         let coin = Coin::<Rng>::new(rng, settings.drop_message_probability).unwrap();
         Self {
             coin,
             stream,
             scheduler,
-            _mix_message: Default::default(),
+            drop_message,
         }
     }
 }
 
-impl<S, Rng, M, Scheduler> Stream for PersistentTransmissionStream<S, Rng, M, Scheduler>
+impl<S, Rng, Scheduler> Stream for PersistentTransmissionStream<S, Rng, Scheduler>
 where
     S: Stream<Item = Vec<u8>> + Unpin,
     Rng: RngCore + Unpin,
-    M: MixMessage + Unpin,
     Scheduler: Stream<Item = ()> + Unpin,
 {
     type Item = Vec<u8>;
@@ -73,6 +69,7 @@ where
             ref mut scheduler,
             ref mut stream,
             ref mut coin,
+            ref drop_message,
             ..
         } = self.get_mut();
         if pin!(scheduler).poll_next_unpin(cx).is_pending() {
@@ -81,17 +78,16 @@ where
         if let Poll::Ready(Some(item)) = pin!(stream).poll_next(cx) {
             Poll::Ready(Some(item))
         } else if coin.flip() {
-            Poll::Ready(Some(M::DROP_MESSAGE.to_vec()))
+            Poll::Ready(Some(drop_message.clone()))
         } else {
             Poll::Pending
         }
     }
 }
 
-pub trait PersistentTransmissionExt<Rng, M, Scheduler>: Stream
+pub trait PersistentTransmissionExt<Rng, Scheduler>: Stream
 where
     Rng: RngCore,
-    M: MixMessage,
     Scheduler: Stream<Item = ()>,
 {
     fn persistent_transmission(
@@ -99,20 +95,19 @@ where
         settings: PersistentTransmissionSettings,
         rng: Rng,
         scheduler: Scheduler,
-    ) -> PersistentTransmissionStream<Self, Rng, M, Scheduler>
+        drop_message: Vec<u8>,
+    ) -> PersistentTransmissionStream<Self, Rng, Scheduler>
     where
         Self: Sized + Unpin,
     {
-        PersistentTransmissionStream::new(settings, self, scheduler, rng)
+        PersistentTransmissionStream::new(settings, self, scheduler, rng, drop_message)
     }
 }
 
-impl<S, Rng, M, Scheduler> PersistentTransmissionExt<Rng, M, Scheduler> for S
+impl<S, Rng, Scheduler> PersistentTransmissionExt<Rng, Scheduler> for S
 where
     S: Stream,
     Rng: RngCore,
-    M: MixMessage,
-    M::PublicKey: Clone + Serialize + DeserializeOwned,
     Scheduler: Stream<Item = ()>,
 {
 }
@@ -150,7 +145,6 @@ enum CoinError {
 mod tests {
     use super::*;
     use futures::StreamExt;
-    use nomos_mix_message::mock::MockMixMessage;
     use rand::SeedableRng;
     use rand_chacha::ChaCha8Rng;
     use std::time::Duration;
@@ -196,16 +190,14 @@ mod tests {
         let lower_bound = expected_emission_interval - torelance;
         let upper_bound = expected_emission_interval + torelance;
         // prepare stream
-        let mut persistent_transmission_stream: PersistentTransmissionStream<
-            _,
-            _,
-            MockMixMessage,
-            _,
-        > = stream.persistent_transmission(
-            settings,
-            ChaCha8Rng::from_entropy(),
-            IntervalStream::new(time::interval(expected_emission_interval)).map(|_| ()),
-        );
+        let drop_message = vec![0u8; 10];
+        let mut persistent_transmission_stream: PersistentTransmissionStream<_, _, _> = stream
+            .persistent_transmission(
+                settings,
+                ChaCha8Rng::from_entropy(),
+                IntervalStream::new(time::interval(expected_emission_interval)).map(|_| ()),
+                drop_message.clone(),
+            );
         // Messages must be scheduled in non-blocking manner.
         schedule_sender.send(vec![1]).unwrap();
         schedule_sender.send(vec![2]).unwrap();
@@ -230,14 +222,16 @@ mod tests {
         );
         assert_interval!(&mut last_time, lower_bound, upper_bound);
 
-        assert!(MockMixMessage::is_drop_message(
-            &persistent_transmission_stream.next().await.unwrap()
-        ));
+        assert_eq!(
+            &persistent_transmission_stream.next().await.unwrap(),
+            &drop_message
+        );
         assert_interval!(&mut last_time, lower_bound, upper_bound);
 
-        assert!(MockMixMessage::is_drop_message(
-            &persistent_transmission_stream.next().await.unwrap()
-        ));
+        assert_eq!(
+            &persistent_transmission_stream.next().await.unwrap(),
+            &drop_message
+        );
         assert_interval!(&mut last_time, lower_bound, upper_bound);
 
         // Schedule a new message and check if it is emitted at the next interval
diff --git a/nomos-mix/message/Cargo.toml b/nomos-mix/message/Cargo.toml
index 3a4c33c1..e6078081 100644
--- a/nomos-mix/message/Cargo.toml
+++ b/nomos-mix/message/Cargo.toml
@@ -6,6 +6,7 @@ edition = "2021"
 [dependencies]
 itertools = "0.13"
 rand_chacha = "0.3"
+serde = { version = "1.0.215", features = ["derive"] }
 sha2 = "0.10"
 sphinx-packet = "0.2"
 thiserror = "1.0.65"
diff --git a/nomos-mix/message/src/lib.rs b/nomos-mix/message/src/lib.rs
index cc8e4c27..b6acd2c2 100644
--- a/nomos-mix/message/src/lib.rs
+++ b/nomos-mix/message/src/lib.rs
@@ -4,10 +4,13 @@ pub mod sphinx;
 pub trait MixMessage {
     type PublicKey;
     type PrivateKey;
+    type Settings;
     type Error;
-    const DROP_MESSAGE: &'static [u8];
+
+    fn new(settings: Self::Settings) -> Self;
 
     fn build_message(
+        &self,
         payload: &[u8],
         public_keys: &[Self::PublicKey],
     ) -> Result<Vec<u8>, Self::Error>;
@@ -19,10 +22,14 @@ pub trait MixMessage {
     /// If the input message was already fully unwrapped, or if its format is invalid,
     /// this function returns `[Error::InvalidMixMessage]`.
     fn unwrap_message(
+        &self,
         message: &[u8],
         private_key: &Self::PrivateKey,
     ) -> Result<(Vec<u8>, bool), Self::Error>;
-    fn is_drop_message(message: &[u8]) -> bool {
-        message == Self::DROP_MESSAGE
+
+    fn drop_message(&self) -> &[u8];
+
+    fn is_drop_message(&self, message: &[u8]) -> bool {
+        message == self.drop_message()
     }
 }
diff --git a/nomos-mix/message/src/mock/mod.rs b/nomos-mix/message/src/mock/mod.rs
index fcb8cf6f..8797ad36 100644
--- a/nomos-mix/message/src/mock/mod.rs
+++ b/nomos-mix/message/src/mock/mod.rs
@@ -1,68 +1,99 @@
 pub mod error;
 
 use error::Error;
+use serde::{Deserialize, Serialize};
 
 use crate::MixMessage;
-// TODO: Remove all the mock below once the actual implementation is integrated to the system.
-//
+
 /// A mock implementation of the Sphinx encoding.
 
 const NODE_ID_SIZE: usize = 32;
-
-// TODO: Move MAX_PAYLOAD_SIZE and MAX_LAYERS to the upper layer (service layer).
-const MAX_PAYLOAD_SIZE: usize = 2048;
 const PAYLOAD_PADDING_SEPARATOR: u8 = 0x01;
-const PAYLOAD_PADDING_SEPARATOR_SIZE: usize = 1;
-const MAX_LAYERS: usize = 5;
-pub const MESSAGE_SIZE: usize =
-    NODE_ID_SIZE * MAX_LAYERS + MAX_PAYLOAD_SIZE + PAYLOAD_PADDING_SEPARATOR_SIZE;
+const PAYLOAD_PADDING_SEPARATOR_SIZE: usize = std::mem::size_of::<u8>();
 
 #[derive(Clone, Debug)]
-pub struct MockMixMessage;
+pub struct MockMixMessage {
+    settings: MockMixMessageSettings,
+    drop_message: Vec<u8>,
+}
+
+#[derive(Clone, Debug, Serialize, Deserialize)]
+pub struct MockMixMessageSettings {
+    max_layers: usize,
+    max_payload_size: usize,
+}
+
+impl MockMixMessageSettings {
+    const fn message_size(&self) -> usize {
+        NODE_ID_SIZE * self.max_layers + self.max_payload_size + PAYLOAD_PADDING_SEPARATOR_SIZE
+    }
+}
+
+impl MockMixMessage {
+    const fn size(&self) -> usize {
+        self.settings.message_size()
+    }
+}
 
 impl MixMessage for MockMixMessage {
     type PublicKey = [u8; NODE_ID_SIZE];
     type PrivateKey = [u8; NODE_ID_SIZE];
+    type Settings = MockMixMessageSettings;
     type Error = Error;
-    const DROP_MESSAGE: &'static [u8] = &[0; MESSAGE_SIZE];
+
+    fn new(settings: Self::Settings) -> Self {
+        let drop_message = vec![0; settings.message_size()];
+        Self {
+            settings,
+            drop_message,
+        }
+    }
 
     /// The length of the encoded message is fixed to [`MESSAGE_SIZE`] bytes.
     /// The [`MAX_LAYERS`] number of [`NodeId`]s are concatenated in front of the payload.
     /// The payload is zero-padded to the end.
-    ///
     fn build_message(
+        &self,
         payload: &[u8],
         public_keys: &[Self::PublicKey],
     ) -> Result<Vec<u8>, Self::Error> {
+        let MockMixMessageSettings {
+            max_layers,
+            max_payload_size,
+        } = self.settings;
+
         // In this mock, we don't encrypt anything. So, we use public key as just a node ID.
         let node_ids = public_keys;
-        if node_ids.is_empty() || node_ids.len() > MAX_LAYERS {
+        if node_ids.is_empty() || node_ids.len() > max_layers {
             return Err(Error::InvalidNumberOfLayers);
         }
-        if payload.len() > MAX_PAYLOAD_SIZE {
+        if payload.len() > max_payload_size {
             return Err(Error::PayloadTooLarge);
         }
 
-        let mut message: Vec<u8> = Vec::with_capacity(MESSAGE_SIZE);
+        let mut message: Vec<u8> = Vec::with_capacity(self.size());
 
         node_ids.iter().for_each(|node_id| {
             message.extend(node_id);
         });
         // If there is any remaining layers, fill them with zeros.
-        (0..MAX_LAYERS - node_ids.len()).for_each(|_| message.extend(&[0; NODE_ID_SIZE]));
+        (0..max_layers - node_ids.len()).for_each(|_| message.extend(&[0; NODE_ID_SIZE]));
 
         // Append payload with padding
         message.extend(payload);
         message.push(PAYLOAD_PADDING_SEPARATOR);
-        message.extend(std::iter::repeat(0).take(MAX_PAYLOAD_SIZE - payload.len()));
+        message.extend(std::iter::repeat(0).take(max_payload_size - payload.len()));
         Ok(message)
     }
 
     fn unwrap_message(
+        &self,
         message: &[u8],
         private_key: &Self::PrivateKey,
     ) -> Result<(Vec<u8>, bool), Self::Error> {
-        if message.len() != MESSAGE_SIZE {
+        let MockMixMessageSettings { max_layers, .. } = self.settings;
+
+        if message.len() != self.size() {
             return Err(Error::InvalidMixMessage);
         }
 
@@ -74,7 +105,7 @@ impl MixMessage for MockMixMessage {
 
         // If this is the last layer
         if message[NODE_ID_SIZE..NODE_ID_SIZE * 2] == [0; NODE_ID_SIZE] {
-            let padded_payload = &message[NODE_ID_SIZE * MAX_LAYERS..];
+            let padded_payload = &message[NODE_ID_SIZE * max_layers..];
             // remove the payload padding
             match padded_payload
                 .iter()
@@ -87,12 +118,16 @@ impl MixMessage for MockMixMessage {
             }
         }
 
-        let mut new_message: Vec<u8> = Vec::with_capacity(MESSAGE_SIZE);
-        new_message.extend(&message[NODE_ID_SIZE..NODE_ID_SIZE * MAX_LAYERS]);
+        let mut new_message: Vec<u8> = Vec::with_capacity(self.size());
+        new_message.extend(&message[NODE_ID_SIZE..NODE_ID_SIZE * max_layers]);
         new_message.extend(&[0; NODE_ID_SIZE]);
-        new_message.extend(&message[NODE_ID_SIZE * MAX_LAYERS..]); // padded payload
+        new_message.extend(&message[NODE_ID_SIZE * max_layers..]); // padded payload
         Ok((new_message, false))
     }
+
+    fn drop_message(&self) -> &[u8] {
+        &self.drop_message
+    }
 }
 
 #[cfg(test)]
@@ -101,23 +136,28 @@ mod tests {
 
     #[test]
     fn message() {
+        let mock_message = MockMixMessage::new(MockMixMessageSettings {
+            max_layers: 5,
+            max_payload_size: 100,
+        });
+
         let node_ids = (0..3).map(|i| [i; NODE_ID_SIZE]).collect::<Vec<_>>();
         let payload = [7; 10];
-        let message = MockMixMessage::build_message(&payload, &node_ids).unwrap();
-        assert_eq!(message.len(), MESSAGE_SIZE);
+        let message = mock_message.build_message(&payload, &node_ids).unwrap();
+        assert_eq!(message.len(), mock_message.size());
 
         let (message, is_fully_unwrapped) =
-            MockMixMessage::unwrap_message(&message, &node_ids[0]).unwrap();
+            mock_message.unwrap_message(&message, &node_ids[0]).unwrap();
         assert!(!is_fully_unwrapped);
-        assert_eq!(message.len(), MESSAGE_SIZE);
+        assert_eq!(message.len(), mock_message.size());
 
         let (message, is_fully_unwrapped) =
-            MockMixMessage::unwrap_message(&message, &node_ids[1]).unwrap();
+            mock_message.unwrap_message(&message, &node_ids[1]).unwrap();
         assert!(!is_fully_unwrapped);
-        assert_eq!(message.len(), MESSAGE_SIZE);
+        assert_eq!(message.len(), mock_message.size());
 
         let (unwrapped_payload, is_fully_unwrapped) =
-            MockMixMessage::unwrap_message(&message, &node_ids[2]).unwrap();
+            mock_message.unwrap_message(&message, &node_ids[2]).unwrap();
         assert!(is_fully_unwrapped);
         assert_eq!(unwrapped_payload, payload);
     }
diff --git a/nomos-mix/message/src/sphinx/mod.rs b/nomos-mix/message/src/sphinx/mod.rs
index c7d62fb5..82b97517 100644
--- a/nomos-mix/message/src/sphinx/mod.rs
+++ b/nomos-mix/message/src/sphinx/mod.rs
@@ -1,5 +1,6 @@
 use error::Error;
 use packet::{Packet, UnpackedPacket};
+use serde::{Deserialize, Serialize};
 
 use crate::MixMessage;
 
@@ -8,21 +9,35 @@ mod layered_cipher;
 pub mod packet;
 mod routing;
 
-pub struct SphinxMessage;
+pub struct SphinxMessage {
+    settings: SphinxMessageSettings,
+    drop_message: Vec<u8>,
+}
+
+#[derive(Clone, Debug, Serialize, Deserialize)]
+pub struct SphinxMessageSettings {
+    max_layers: usize,
+    max_payload_size: usize,
+}
 
 const ASYM_KEY_SIZE: usize = 32;
-// TODO: Move these constants to the upper layer (service layer).
-const MAX_PAYLOAD_SIZE: usize = 2048;
-const MAX_LAYERS: usize = 5;
 
 impl MixMessage for SphinxMessage {
     type PublicKey = [u8; ASYM_KEY_SIZE];
     type PrivateKey = [u8; ASYM_KEY_SIZE];
+    type Settings = SphinxMessageSettings;
     type Error = Error;
 
-    const DROP_MESSAGE: &'static [u8] = &[0; Packet::size(MAX_LAYERS, MAX_PAYLOAD_SIZE)];
+    fn new(settings: Self::Settings) -> Self {
+        let drop_message = vec![0; Packet::size(settings.max_layers, settings.max_payload_size)];
+        Self {
+            settings,
+            drop_message,
+        }
+    }
 
     fn build_message(
+        &self,
         payload: &[u8],
         public_keys: &[Self::PublicKey],
     ) -> Result<Vec<u8>, Self::Error> {
@@ -31,25 +46,32 @@ impl MixMessage for SphinxMessage {
                 .iter()
                 .map(|k| x25519_dalek::PublicKey::from(*k))
                 .collect::<Vec<_>>(),
-            MAX_LAYERS,
+            self.settings.max_layers,
             payload,
-            MAX_PAYLOAD_SIZE,
+            self.settings.max_payload_size,
         )?;
         Ok(packet.to_bytes())
     }
 
     fn unwrap_message(
+        &self,
         message: &[u8],
         private_key: &Self::PrivateKey,
     ) -> Result<(Vec<u8>, bool), Self::Error> {
-        let packet = Packet::from_bytes(message, MAX_LAYERS)?;
-        let unpacked_packet =
-            packet.unpack(&x25519_dalek::StaticSecret::from(*private_key), MAX_LAYERS)?;
+        let packet = Packet::from_bytes(message, self.settings.max_layers)?;
+        let unpacked_packet = packet.unpack(
+            &x25519_dalek::StaticSecret::from(*private_key),
+            self.settings.max_layers,
+        )?;
         match unpacked_packet {
             UnpackedPacket::ToForward(packet) => Ok((packet.to_bytes(), false)),
             UnpackedPacket::FullyUnpacked(payload) => Ok((payload, true)),
         }
     }
+
+    fn drop_message(&self) -> &[u8] {
+        &self.drop_message
+    }
 }
 
 fn parse_bytes<'a>(data: &'a [u8], sizes: &[usize]) -> Result<Vec<&'a [u8]>, String> {