diff --git a/CODEOWNERS b/CODEOWNERS
index 0174f633cd..d07d00632c 100644
--- a/CODEOWNERS
+++ b/CODEOWNERS
@@ -111,6 +111,8 @@ esphome/components/hte501/* @Stock-M
 esphome/components/hydreon_rgxx/* @functionpointer
 esphome/components/i2c/* @esphome/core
 esphome/components/i2s_audio/* @jesserockz
+esphome/components/i2s_audio/media_player/* @jesserockz
+esphome/components/i2s_audio/microphone/* @jesserockz
 esphome/components/ili9xxx/* @nielsnl68
 esphome/components/improv_base/* @esphome/core
 esphome/components/improv_serial/* @esphome/core
@@ -154,6 +156,7 @@ esphome/components/mcp9808/* @k7hpn
 esphome/components/md5/* @esphome/core
 esphome/components/mdns/* @esphome/core
 esphome/components/media_player/* @jesserockz
+esphome/components/microphone/* @jesserockz
 esphome/components/mics_4514/* @jesserockz
 esphome/components/midea/* @dudanov
 esphome/components/midea_ir/* @dudanov
@@ -288,6 +291,7 @@ esphome/components/ufire_ise/* @pvizeli
 esphome/components/ultrasonic/* @OttoWinter
 esphome/components/vbus/* @ssieb
 esphome/components/version/* @esphome/core
+esphome/components/voice_assistant/* @jesserockz
 esphome/components/wake_on_lan/* @willwill2will54
 esphome/components/web_server_base/* @OttoWinter
 esphome/components/whirlpool/* @glmnet
diff --git a/esphome/components/api/api.proto b/esphome/components/api/api.proto
index e36f0581ca..f31ef3ffc0 100644
--- a/esphome/components/api/api.proto
+++ b/esphome/components/api/api.proto
@@ -55,6 +55,7 @@ service APIConnection {
   rpc subscribe_bluetooth_connections_free(SubscribeBluetoothConnectionsFreeRequest) returns (BluetoothConnectionsFreeResponse) {}
   rpc unsubscribe_bluetooth_le_advertisements(UnsubscribeBluetoothLEAdvertisementsRequest) returns (void) {}
 
+  rpc subscribe_voice_assistant(SubscribeVoiceAssistantRequest) returns (void) {}
 }
 
 
@@ -210,6 +211,8 @@ message DeviceInfoResponse {
   string manufacturer = 12;
 
   string friendly_name = 13;
+
+  uint32 voice_assistant_version = 14;
 }
 
 message ListEntitiesRequest {
@@ -1379,3 +1382,55 @@ message BluetoothDeviceClearCacheResponse {
   bool success = 2;
   int32 error = 3;
 }
+
+// ==================== PUSH TO TALK ====================
+message SubscribeVoiceAssistantRequest {
+  option (id) = 89;
+  option (source) = SOURCE_CLIENT;
+  option (ifdef) = "USE_VOICE_ASSISTANT";
+
+  bool subscribe = 1;
+}
+
+message VoiceAssistantRequest {
+  option (id) = 90;
+  option (source) = SOURCE_SERVER;
+  option (ifdef) = "USE_VOICE_ASSISTANT";
+
+  bool start = 1;
+}
+
+message VoiceAssistantResponse {
+  option (id) = 91;
+  option (source) = SOURCE_CLIENT;
+  option (ifdef) = "USE_VOICE_ASSISTANT";
+
+  uint32 port = 1;
+  bool error = 2;
+}
+
+enum VoiceAssistantEvent {
+  VOICE_ASSISTANT_ERROR = 0;
+  VOICE_ASSISTANT_RUN_START = 1;
+  VOICE_ASSISTANT_RUN_END = 2;
+  VOICE_ASSISTANT_STT_START = 3;
+  VOICE_ASSISTANT_STT_END = 4;
+  VOICE_ASSISTANT_INTENT_START = 5;
+  VOICE_ASSISTANT_INTENT_END = 6;
+  VOICE_ASSISTANT_TTS_START = 7;
+  VOICE_ASSISTANT_TTS_END = 8;
+}
+
+message VoiceAssistantEventData {
+  string name = 1;
+  string value = 2;
+}
+
+message VoiceAssistantEventResponse {
+  option (id) = 92;
+  option (source) = SOURCE_CLIENT;
+  option (ifdef) = "USE_VOICE_ASSISTANT";
+
+  VoiceAssistantEvent event_type = 1;
+  repeated VoiceAssistantEventData data = 2;
+}
diff --git a/esphome/components/api/api_connection.cpp b/esphome/components/api/api_connection.cpp
index 104560771e..96fb3ea9fa 100644
--- a/esphome/components/api/api_connection.cpp
+++ b/esphome/components/api/api_connection.cpp
@@ -16,6 +16,9 @@
 #ifdef USE_BLUETOOTH_PROXY
 #include "esphome/components/bluetooth_proxy/bluetooth_proxy.h"
 #endif
+#ifdef USE_VOICE_ASSISTANT
+#include "esphome/components/voice_assistant/voice_assistant.h"
+#endif
 
 namespace esphome {
 namespace api {
@@ -893,6 +896,30 @@ BluetoothConnectionsFreeResponse APIConnection::subscribe_bluetooth_connections_
 }
 #endif
 
+#ifdef USE_VOICE_ASSISTANT
+bool APIConnection::request_voice_assistant(bool start) {
+  if (!this->voice_assistant_subscription_)
+    return false;
+  VoiceAssistantRequest msg;
+  msg.start = start;
+  return this->send_voice_assistant_request(msg);
+}
+void APIConnection::on_voice_assistant_response(const VoiceAssistantResponse &msg) {
+  if (voice_assistant::global_voice_assistant != nullptr) {
+    struct sockaddr_storage storage;
+    socklen_t len = sizeof(storage);
+    this->helper_->getpeername((struct sockaddr *) &storage, &len);
+    voice_assistant::global_voice_assistant->start(&storage, msg.port);
+  }
+};
+void APIConnection::on_voice_assistant_event_response(const VoiceAssistantEventResponse &msg) {
+  if (voice_assistant::global_voice_assistant != nullptr) {
+    voice_assistant::global_voice_assistant->on_event(msg);
+  }
+}
+
+#endif
+
 bool APIConnection::send_log_message(int level, const char *tag, const char *line) {
   if (this->log_subscription_ < level)
     return false;
@@ -970,6 +997,9 @@ DeviceInfoResponse APIConnection::device_info(const DeviceInfoRequest &msg) {
   resp.bluetooth_proxy_version = bluetooth_proxy::global_bluetooth_proxy->has_active()
                                      ? bluetooth_proxy::ACTIVE_CONNECTIONS_VERSION
                                      : bluetooth_proxy::PASSIVE_ONLY_VERSION;
+#endif
+#ifdef USE_VOICE_ASSISTANT
+  resp.voice_assistant_version = 1;
 #endif
   return resp;
 }
diff --git a/esphome/components/api/api_connection.h b/esphome/components/api/api_connection.h
index c85c69a2b9..78ecbb98e6 100644
--- a/esphome/components/api/api_connection.h
+++ b/esphome/components/api/api_connection.h
@@ -6,6 +6,7 @@
 #include "api_server.h"
 #include "esphome/core/application.h"
 #include "esphome/core/component.h"
+#include "esphome/core/defines.h"
 
 #include <vector>
 
@@ -123,6 +124,15 @@ class APIConnection : public APIServerConnection {
   }
 #endif
 
+#ifdef USE_VOICE_ASSISTANT
+  void subscribe_voice_assistant(const SubscribeVoiceAssistantRequest &msg) override {
+    this->voice_assistant_subscription_ = msg.subscribe;
+  }
+  bool request_voice_assistant(bool start);
+  void on_voice_assistant_response(const VoiceAssistantResponse &msg) override;
+  void on_voice_assistant_event_response(const VoiceAssistantEventResponse &msg) override;
+#endif
+
   void on_disconnect_response(const DisconnectResponse &value) override;
   void on_ping_response(const PingResponse &value) override {
     // we initiated ping
@@ -203,6 +213,9 @@ class APIConnection : public APIServerConnection {
   bool service_call_subscription_{false};
 #ifdef USE_BLUETOOTH_PROXY
   bool bluetooth_le_advertisement_subscription_{false};
+#endif
+#ifdef USE_VOICE_ASSISTANT
+  bool voice_assistant_subscription_{false};
 #endif
   bool next_close_ = false;
   APIServer *parent_;
diff --git a/esphome/components/api/api_frame_helper.h b/esphome/components/api/api_frame_helper.h
index 348a9b574f..bf4872d2d6 100644
--- a/esphome/components/api/api_frame_helper.h
+++ b/esphome/components/api/api_frame_helper.h
@@ -10,8 +10,8 @@
 #include "noise/protocol.h"
 #endif
 
-#include "esphome/components/socket/socket.h"
 #include "api_noise_context.h"
+#include "esphome/components/socket/socket.h"
 
 namespace esphome {
 namespace api {
@@ -67,6 +67,7 @@ class APIFrameHelper {
   virtual bool can_write_without_blocking() = 0;
   virtual APIError write_packet(uint16_t type, const uint8_t *data, size_t len) = 0;
   virtual std::string getpeername() = 0;
+  virtual int getpeername(struct sockaddr *addr, socklen_t *addrlen) = 0;
   virtual APIError close() = 0;
   virtual APIError shutdown(int how) = 0;
   // Give this helper a name for logging
@@ -84,7 +85,10 @@ class APINoiseFrameHelper : public APIFrameHelper {
   APIError read_packet(ReadPacketBuffer *buffer) override;
   bool can_write_without_blocking() override;
   APIError write_packet(uint16_t type, const uint8_t *payload, size_t len) override;
-  std::string getpeername() override { return socket_->getpeername(); }
+  std::string getpeername() override { return this->socket_->getpeername(); }
+  int getpeername(struct sockaddr *addr, socklen_t *addrlen) override {
+    return this->socket_->getpeername(addr, addrlen);
+  }
   APIError close() override;
   APIError shutdown(int how) override;
   // Give this helper a name for logging
@@ -144,7 +148,10 @@ class APIPlaintextFrameHelper : public APIFrameHelper {
   APIError read_packet(ReadPacketBuffer *buffer) override;
   bool can_write_without_blocking() override;
   APIError write_packet(uint16_t type, const uint8_t *payload, size_t len) override;
-  std::string getpeername() override { return socket_->getpeername(); }
+  std::string getpeername() override { return this->socket_->getpeername(); }
+  int getpeername(struct sockaddr *addr, socklen_t *addrlen) override {
+    return this->socket_->getpeername(addr, addrlen);
+  }
   APIError close() override;
   APIError shutdown(int how) override;
   // Give this helper a name for logging
diff --git a/esphome/components/api/api_pb2.cpp b/esphome/components/api/api_pb2.cpp
index 6260020064..334cde16b3 100644
--- a/esphome/components/api/api_pb2.cpp
+++ b/esphome/components/api/api_pb2.cpp
@@ -407,6 +407,32 @@ const char *proto_enum_to_string<enums::BluetoothDeviceRequestType>(enums::Bluet
   }
 }
 #endif
+#ifdef HAS_PROTO_MESSAGE_DUMP
+template<> const char *proto_enum_to_string<enums::VoiceAssistantEvent>(enums::VoiceAssistantEvent value) {
+  switch (value) {
+    case enums::VOICE_ASSISTANT_ERROR:
+      return "VOICE_ASSISTANT_ERROR";
+    case enums::VOICE_ASSISTANT_RUN_START:
+      return "VOICE_ASSISTANT_RUN_START";
+    case enums::VOICE_ASSISTANT_RUN_END:
+      return "VOICE_ASSISTANT_RUN_END";
+    case enums::VOICE_ASSISTANT_STT_START:
+      return "VOICE_ASSISTANT_STT_START";
+    case enums::VOICE_ASSISTANT_STT_END:
+      return "VOICE_ASSISTANT_STT_END";
+    case enums::VOICE_ASSISTANT_INTENT_START:
+      return "VOICE_ASSISTANT_INTENT_START";
+    case enums::VOICE_ASSISTANT_INTENT_END:
+      return "VOICE_ASSISTANT_INTENT_END";
+    case enums::VOICE_ASSISTANT_TTS_START:
+      return "VOICE_ASSISTANT_TTS_START";
+    case enums::VOICE_ASSISTANT_TTS_END:
+      return "VOICE_ASSISTANT_TTS_END";
+    default:
+      return "UNKNOWN";
+  }
+}
+#endif
 bool HelloRequest::decode_varint(uint32_t field_id, ProtoVarInt value) {
   switch (field_id) {
     case 2: {
@@ -594,6 +620,10 @@ bool DeviceInfoResponse::decode_varint(uint32_t field_id, ProtoVarInt value) {
       this->bluetooth_proxy_version = value.as_uint32();
       return true;
     }
+    case 14: {
+      this->voice_assistant_version = value.as_uint32();
+      return true;
+    }
     default:
       return false;
   }
@@ -654,6 +684,7 @@ void DeviceInfoResponse::encode(ProtoWriteBuffer buffer) const {
   buffer.encode_uint32(11, this->bluetooth_proxy_version);
   buffer.encode_string(12, this->manufacturer);
   buffer.encode_string(13, this->friendly_name);
+  buffer.encode_uint32(14, this->voice_assistant_version);
 }
 #ifdef HAS_PROTO_MESSAGE_DUMP
 void DeviceInfoResponse::dump_to(std::string &out) const {
@@ -712,6 +743,11 @@ void DeviceInfoResponse::dump_to(std::string &out) const {
   out.append("  friendly_name: ");
   out.append("'").append(this->friendly_name).append("'");
   out.append("\n");
+
+  out.append("  voice_assistant_version: ");
+  sprintf(buffer, "%u", this->voice_assistant_version);
+  out.append(buffer);
+  out.append("\n");
   out.append("}");
 }
 #endif
@@ -6111,6 +6147,155 @@ void BluetoothDeviceClearCacheResponse::dump_to(std::string &out) const {
   out.append("}");
 }
 #endif
+bool SubscribeVoiceAssistantRequest::decode_varint(uint32_t field_id, ProtoVarInt value) {
+  switch (field_id) {
+    case 1: {
+      this->subscribe = value.as_bool();
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+void SubscribeVoiceAssistantRequest::encode(ProtoWriteBuffer buffer) const { buffer.encode_bool(1, this->subscribe); }
+#ifdef HAS_PROTO_MESSAGE_DUMP
+void SubscribeVoiceAssistantRequest::dump_to(std::string &out) const {
+  __attribute__((unused)) char buffer[64];
+  out.append("SubscribeVoiceAssistantRequest {\n");
+  out.append("  subscribe: ");
+  out.append(YESNO(this->subscribe));
+  out.append("\n");
+  out.append("}");
+}
+#endif
+bool VoiceAssistantRequest::decode_varint(uint32_t field_id, ProtoVarInt value) {
+  switch (field_id) {
+    case 1: {
+      this->start = value.as_bool();
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+void VoiceAssistantRequest::encode(ProtoWriteBuffer buffer) const { buffer.encode_bool(1, this->start); }
+#ifdef HAS_PROTO_MESSAGE_DUMP
+void VoiceAssistantRequest::dump_to(std::string &out) const {
+  __attribute__((unused)) char buffer[64];
+  out.append("VoiceAssistantRequest {\n");
+  out.append("  start: ");
+  out.append(YESNO(this->start));
+  out.append("\n");
+  out.append("}");
+}
+#endif
+bool VoiceAssistantResponse::decode_varint(uint32_t field_id, ProtoVarInt value) {
+  switch (field_id) {
+    case 1: {
+      this->port = value.as_uint32();
+      return true;
+    }
+    case 2: {
+      this->error = value.as_bool();
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+void VoiceAssistantResponse::encode(ProtoWriteBuffer buffer) const {
+  buffer.encode_uint32(1, this->port);
+  buffer.encode_bool(2, this->error);
+}
+#ifdef HAS_PROTO_MESSAGE_DUMP
+void VoiceAssistantResponse::dump_to(std::string &out) const {
+  __attribute__((unused)) char buffer[64];
+  out.append("VoiceAssistantResponse {\n");
+  out.append("  port: ");
+  sprintf(buffer, "%u", this->port);
+  out.append(buffer);
+  out.append("\n");
+
+  out.append("  error: ");
+  out.append(YESNO(this->error));
+  out.append("\n");
+  out.append("}");
+}
+#endif
+bool VoiceAssistantEventData::decode_length(uint32_t field_id, ProtoLengthDelimited value) {
+  switch (field_id) {
+    case 1: {
+      this->name = value.as_string();
+      return true;
+    }
+    case 2: {
+      this->value = value.as_string();
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+void VoiceAssistantEventData::encode(ProtoWriteBuffer buffer) const {
+  buffer.encode_string(1, this->name);
+  buffer.encode_string(2, this->value);
+}
+#ifdef HAS_PROTO_MESSAGE_DUMP
+void VoiceAssistantEventData::dump_to(std::string &out) const {
+  __attribute__((unused)) char buffer[64];
+  out.append("VoiceAssistantEventData {\n");
+  out.append("  name: ");
+  out.append("'").append(this->name).append("'");
+  out.append("\n");
+
+  out.append("  value: ");
+  out.append("'").append(this->value).append("'");
+  out.append("\n");
+  out.append("}");
+}
+#endif
+bool VoiceAssistantEventResponse::decode_varint(uint32_t field_id, ProtoVarInt value) {
+  switch (field_id) {
+    case 1: {
+      this->event_type = value.as_enum<enums::VoiceAssistantEvent>();
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+bool VoiceAssistantEventResponse::decode_length(uint32_t field_id, ProtoLengthDelimited value) {
+  switch (field_id) {
+    case 2: {
+      this->data.push_back(value.as_message<VoiceAssistantEventData>());
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+void VoiceAssistantEventResponse::encode(ProtoWriteBuffer buffer) const {
+  buffer.encode_enum<enums::VoiceAssistantEvent>(1, this->event_type);
+  for (auto &it : this->data) {
+    buffer.encode_message<VoiceAssistantEventData>(2, it, true);
+  }
+}
+#ifdef HAS_PROTO_MESSAGE_DUMP
+void VoiceAssistantEventResponse::dump_to(std::string &out) const {
+  __attribute__((unused)) char buffer[64];
+  out.append("VoiceAssistantEventResponse {\n");
+  out.append("  event_type: ");
+  out.append(proto_enum_to_string<enums::VoiceAssistantEvent>(this->event_type));
+  out.append("\n");
+
+  for (const auto &it : this->data) {
+    out.append("  data: ");
+    it.dump_to(out);
+    out.append("\n");
+  }
+  out.append("}");
+}
+#endif
 
 }  // namespace api
 }  // namespace esphome
diff --git a/esphome/components/api/api_pb2.h b/esphome/components/api/api_pb2.h
index ade9b9cc8f..9f71c07913 100644
--- a/esphome/components/api/api_pb2.h
+++ b/esphome/components/api/api_pb2.h
@@ -165,6 +165,17 @@ enum BluetoothDeviceRequestType : uint32_t {
   BLUETOOTH_DEVICE_REQUEST_TYPE_CONNECT_V3_WITHOUT_CACHE = 5,
   BLUETOOTH_DEVICE_REQUEST_TYPE_CLEAR_CACHE = 6,
 };
+enum VoiceAssistantEvent : uint32_t {
+  VOICE_ASSISTANT_ERROR = 0,
+  VOICE_ASSISTANT_RUN_START = 1,
+  VOICE_ASSISTANT_RUN_END = 2,
+  VOICE_ASSISTANT_STT_START = 3,
+  VOICE_ASSISTANT_STT_END = 4,
+  VOICE_ASSISTANT_INTENT_START = 5,
+  VOICE_ASSISTANT_INTENT_END = 6,
+  VOICE_ASSISTANT_TTS_START = 7,
+  VOICE_ASSISTANT_TTS_END = 8,
+};
 
 }  // namespace enums
 
@@ -279,6 +290,7 @@ class DeviceInfoResponse : public ProtoMessage {
   uint32_t bluetooth_proxy_version{0};
   std::string manufacturer{};
   std::string friendly_name{};
+  uint32_t voice_assistant_version{0};
   void encode(ProtoWriteBuffer buffer) const override;
 #ifdef HAS_PROTO_MESSAGE_DUMP
   void dump_to(std::string &out) const override;
@@ -1577,6 +1589,65 @@ class BluetoothDeviceClearCacheResponse : public ProtoMessage {
  protected:
   bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
 };
+class SubscribeVoiceAssistantRequest : public ProtoMessage {
+ public:
+  bool subscribe{false};
+  void encode(ProtoWriteBuffer buffer) const override;
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  void dump_to(std::string &out) const override;
+#endif
+
+ protected:
+  bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
+};
+class VoiceAssistantRequest : public ProtoMessage {
+ public:
+  bool start{false};
+  void encode(ProtoWriteBuffer buffer) const override;
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  void dump_to(std::string &out) const override;
+#endif
+
+ protected:
+  bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
+};
+class VoiceAssistantResponse : public ProtoMessage {
+ public:
+  uint32_t port{0};
+  bool error{false};
+  void encode(ProtoWriteBuffer buffer) const override;
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  void dump_to(std::string &out) const override;
+#endif
+
+ protected:
+  bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
+};
+class VoiceAssistantEventData : public ProtoMessage {
+ public:
+  std::string name{};
+  std::string value{};
+  void encode(ProtoWriteBuffer buffer) const override;
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  void dump_to(std::string &out) const override;
+#endif
+
+ protected:
+  bool decode_length(uint32_t field_id, ProtoLengthDelimited value) override;
+};
+class VoiceAssistantEventResponse : public ProtoMessage {
+ public:
+  enums::VoiceAssistantEvent event_type{};
+  std::vector<VoiceAssistantEventData> data{};
+  void encode(ProtoWriteBuffer buffer) const override;
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  void dump_to(std::string &out) const override;
+#endif
+
+ protected:
+  bool decode_length(uint32_t field_id, ProtoLengthDelimited value) override;
+  bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
+};
 
 }  // namespace api
 }  // namespace esphome
diff --git a/esphome/components/api/api_pb2_service.cpp b/esphome/components/api/api_pb2_service.cpp
index 7d019e1d3d..df36d0fdea 100644
--- a/esphome/components/api/api_pb2_service.cpp
+++ b/esphome/components/api/api_pb2_service.cpp
@@ -453,6 +453,20 @@ bool APIServerConnectionBase::send_bluetooth_device_clear_cache_response(const B
   return this->send_message_<BluetoothDeviceClearCacheResponse>(msg, 88);
 }
 #endif
+#ifdef USE_VOICE_ASSISTANT
+#endif
+#ifdef USE_VOICE_ASSISTANT
+bool APIServerConnectionBase::send_voice_assistant_request(const VoiceAssistantRequest &msg) {
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  ESP_LOGVV(TAG, "send_voice_assistant_request: %s", msg.dump().c_str());
+#endif
+  return this->send_message_<VoiceAssistantRequest>(msg, 90);
+}
+#endif
+#ifdef USE_VOICE_ASSISTANT
+#endif
+#ifdef USE_VOICE_ASSISTANT
+#endif
 bool APIServerConnectionBase::read_message(uint32_t msg_size, uint32_t msg_type, uint8_t *msg_data) {
   switch (msg_type) {
     case 1: {
@@ -827,6 +841,39 @@ bool APIServerConnectionBase::read_message(uint32_t msg_size, uint32_t msg_type,
       ESP_LOGVV(TAG, "on_unsubscribe_bluetooth_le_advertisements_request: %s", msg.dump().c_str());
 #endif
       this->on_unsubscribe_bluetooth_le_advertisements_request(msg);
+#endif
+      break;
+    }
+    case 89: {
+#ifdef USE_VOICE_ASSISTANT
+      SubscribeVoiceAssistantRequest msg;
+      msg.decode(msg_data, msg_size);
+#ifdef HAS_PROTO_MESSAGE_DUMP
+      ESP_LOGVV(TAG, "on_subscribe_voice_assistant_request: %s", msg.dump().c_str());
+#endif
+      this->on_subscribe_voice_assistant_request(msg);
+#endif
+      break;
+    }
+    case 91: {
+#ifdef USE_VOICE_ASSISTANT
+      VoiceAssistantResponse msg;
+      msg.decode(msg_data, msg_size);
+#ifdef HAS_PROTO_MESSAGE_DUMP
+      ESP_LOGVV(TAG, "on_voice_assistant_response: %s", msg.dump().c_str());
+#endif
+      this->on_voice_assistant_response(msg);
+#endif
+      break;
+    }
+    case 92: {
+#ifdef USE_VOICE_ASSISTANT
+      VoiceAssistantEventResponse msg;
+      msg.decode(msg_data, msg_size);
+#ifdef HAS_PROTO_MESSAGE_DUMP
+      ESP_LOGVV(TAG, "on_voice_assistant_event_response: %s", msg.dump().c_str());
+#endif
+      this->on_voice_assistant_event_response(msg);
 #endif
       break;
     }
@@ -1226,6 +1273,19 @@ void APIServerConnection::on_unsubscribe_bluetooth_le_advertisements_request(
   this->unsubscribe_bluetooth_le_advertisements(msg);
 }
 #endif
+#ifdef USE_VOICE_ASSISTANT
+void APIServerConnection::on_subscribe_voice_assistant_request(const SubscribeVoiceAssistantRequest &msg) {
+  if (!this->is_connection_setup()) {
+    this->on_no_setup_connection();
+    return;
+  }
+  if (!this->is_authenticated()) {
+    this->on_unauthenticated_access();
+    return;
+  }
+  this->subscribe_voice_assistant(msg);
+}
+#endif
 
 }  // namespace api
 }  // namespace esphome
diff --git a/esphome/components/api/api_pb2_service.h b/esphome/components/api/api_pb2_service.h
index 457a3d28a9..3808f128a4 100644
--- a/esphome/components/api/api_pb2_service.h
+++ b/esphome/components/api/api_pb2_service.h
@@ -224,6 +224,18 @@ class APIServerConnectionBase : public ProtoService {
 #endif
 #ifdef USE_BLUETOOTH_PROXY
   bool send_bluetooth_device_clear_cache_response(const BluetoothDeviceClearCacheResponse &msg);
+#endif
+#ifdef USE_VOICE_ASSISTANT
+  virtual void on_subscribe_voice_assistant_request(const SubscribeVoiceAssistantRequest &value){};
+#endif
+#ifdef USE_VOICE_ASSISTANT
+  bool send_voice_assistant_request(const VoiceAssistantRequest &msg);
+#endif
+#ifdef USE_VOICE_ASSISTANT
+  virtual void on_voice_assistant_response(const VoiceAssistantResponse &value){};
+#endif
+#ifdef USE_VOICE_ASSISTANT
+  virtual void on_voice_assistant_event_response(const VoiceAssistantEventResponse &value){};
 #endif
  protected:
   bool read_message(uint32_t msg_size, uint32_t msg_type, uint8_t *msg_data) override;
@@ -306,6 +318,9 @@ class APIServerConnection : public APIServerConnectionBase {
 #endif
 #ifdef USE_BLUETOOTH_PROXY
   virtual void unsubscribe_bluetooth_le_advertisements(const UnsubscribeBluetoothLEAdvertisementsRequest &msg) = 0;
+#endif
+#ifdef USE_VOICE_ASSISTANT
+  virtual void subscribe_voice_assistant(const SubscribeVoiceAssistantRequest &msg) = 0;
 #endif
  protected:
   void on_hello_request(const HelloRequest &msg) override;
@@ -384,6 +399,9 @@ class APIServerConnection : public APIServerConnectionBase {
   void on_unsubscribe_bluetooth_le_advertisements_request(
       const UnsubscribeBluetoothLEAdvertisementsRequest &msg) override;
 #endif
+#ifdef USE_VOICE_ASSISTANT
+  void on_subscribe_voice_assistant_request(const SubscribeVoiceAssistantRequest &msg) override;
+#endif
 };
 
 }  // namespace api
diff --git a/esphome/components/api/api_server.cpp b/esphome/components/api/api_server.cpp
index c60766b364..fbef4b253f 100644
--- a/esphome/components/api/api_server.cpp
+++ b/esphome/components/api/api_server.cpp
@@ -427,5 +427,18 @@ void APIServer::on_shutdown() {
   delay(10);
 }
 
+#ifdef USE_VOICE_ASSISTANT
+void APIServer::start_voice_assistant() {
+  for (auto &c : this->clients_) {
+    c->request_voice_assistant(true);
+  }
+}
+void APIServer::stop_voice_assistant() {
+  for (auto &c : this->clients_) {
+    c->request_voice_assistant(false);
+  }
+}
+#endif
+
 }  // namespace api
 }  // namespace esphome
diff --git a/esphome/components/api/api_server.h b/esphome/components/api/api_server.h
index db87affdb8..30103b2e3f 100644
--- a/esphome/components/api/api_server.h
+++ b/esphome/components/api/api_server.h
@@ -95,6 +95,11 @@ class APIServer : public Component, public Controller {
   void request_time();
 #endif
 
+#ifdef USE_VOICE_ASSISTANT
+  void start_voice_assistant();
+  void stop_voice_assistant();
+#endif
+
   bool is_connected() const;
 
   struct HomeAssistantStateSubscription {
diff --git a/esphome/components/i2s_audio/__init__.py b/esphome/components/i2s_audio/__init__.py
index e69de29bb2..1773d3082f 100644
--- a/esphome/components/i2s_audio/__init__.py
+++ b/esphome/components/i2s_audio/__init__.py
@@ -0,0 +1,70 @@
+import esphome.config_validation as cv
+import esphome.final_validate as fv
+import esphome.codegen as cg
+
+from esphome import pins
+from esphome.const import CONF_ID
+from esphome.components.esp32 import get_esp32_variant
+from esphome.components.esp32.const import (
+    VARIANT_ESP32,
+    VARIANT_ESP32S2,
+    VARIANT_ESP32S3,
+    VARIANT_ESP32C3,
+)
+
+CODEOWNERS = ["@jesserockz"]
+DEPENDENCIES = ["esp32"]
+MULTI_CONF = True
+
+CONF_I2S_DOUT_PIN = "i2s_dout_pin"
+CONF_I2S_DIN_PIN = "i2s_din_pin"
+CONF_I2S_BCLK_PIN = "i2s_bclk_pin"
+CONF_I2S_LRCLK_PIN = "i2s_lrclk_pin"
+
+CONF_I2S_AUDIO = "i2s_audio"
+CONF_I2S_AUDIO_ID = "i2s_audio_id"
+
+i2s_audio_ns = cg.esphome_ns.namespace("i2s_audio")
+I2SAudioComponent = i2s_audio_ns.class_("I2SAudioComponent", cg.Component)
+I2SAudioIn = i2s_audio_ns.class_("I2SAudioIn", cg.Parented.template(I2SAudioComponent))
+I2SAudioOut = i2s_audio_ns.class_(
+    "I2SAudioOut", cg.Parented.template(I2SAudioComponent)
+)
+
+# https://github.com/espressif/esp-idf/blob/master/components/soc/{variant}/include/soc/soc_caps.h
+I2S_PORTS = {
+    VARIANT_ESP32: 2,
+    VARIANT_ESP32S2: 1,
+    VARIANT_ESP32S3: 2,
+    VARIANT_ESP32C3: 1,
+}
+
+CONFIG_SCHEMA = cv.Schema(
+    {
+        cv.GenerateID(): cv.declare_id(I2SAudioComponent),
+        cv.Required(CONF_I2S_BCLK_PIN): pins.internal_gpio_output_pin_number,
+        cv.Required(CONF_I2S_LRCLK_PIN): pins.internal_gpio_output_pin_number,
+    }
+)
+
+
+def _final_validate(_):
+    i2s_audio_configs = fv.full_config.get()[CONF_I2S_AUDIO]
+    variant = get_esp32_variant()
+    if variant not in I2S_PORTS:
+        raise cv.Invalid(f"Unsupported variant {variant}")
+    if len(i2s_audio_configs) > I2S_PORTS[variant]:
+        raise cv.Invalid(
+            f"Only {I2S_PORTS[variant]} I2S audio ports are supported on {variant}"
+        )
+
+
+FINAL_VALIDATE_SCHEMA = _final_validate
+
+
+async def to_code(config):
+    var = cg.new_Pvariable(config[CONF_ID])
+    await cg.register_component(var, config)
+
+    cg.add(var.set_bclk_pin(config[CONF_I2S_BCLK_PIN]))
+    cg.add(var.set_lrclk_pin(config[CONF_I2S_LRCLK_PIN]))
diff --git a/esphome/components/i2s_audio/i2s_audio.cpp b/esphome/components/i2s_audio/i2s_audio.cpp
new file mode 100644
index 0000000000..c1a608c064
--- /dev/null
+++ b/esphome/components/i2s_audio/i2s_audio.cpp
@@ -0,0 +1,30 @@
+#include "i2s_audio.h"
+
+#ifdef USE_ESP32
+
+#include "esphome/core/log.h"
+
+namespace esphome {
+namespace i2s_audio {
+
+static const char *const TAG = "i2s_audio";
+
+void I2SAudioComponent::setup() {
+  static i2s_port_t next_port_num = I2S_NUM_0;
+
+  if (next_port_num >= I2S_NUM_MAX) {
+    ESP_LOGE(TAG, "Too many I2S Audio components!");
+    this->mark_failed();
+    return;
+  }
+
+  this->port_ = next_port_num;
+  next_port_num = (i2s_port_t) (next_port_num + 1);
+
+  ESP_LOGCONFIG(TAG, "Setting up I2S Audio...");
+}
+
+}  // namespace i2s_audio
+}  // namespace esphome
+
+#endif  // USE_ESP32
diff --git a/esphome/components/i2s_audio/i2s_audio.h b/esphome/components/i2s_audio/i2s_audio.h
new file mode 100644
index 0000000000..6b3fa10f3c
--- /dev/null
+++ b/esphome/components/i2s_audio/i2s_audio.h
@@ -0,0 +1,64 @@
+#pragma once
+
+#ifdef USE_ESP32
+
+#include <driver/i2s.h>
+#include "esphome/core/component.h"
+#include "esphome/core/helpers.h"
+
+namespace esphome {
+namespace i2s_audio {
+
+class I2SAudioComponent;
+
+class I2SAudioIn : public Parented<I2SAudioComponent> {};
+
+class I2SAudioOut : public Parented<I2SAudioComponent> {};
+
+class I2SAudioComponent : public Component {
+ public:
+  void setup() override;
+
+  void register_audio_in(I2SAudioIn *in) {
+    this->audio_in_ = in;
+    in->set_parent(this);
+  }
+  void register_audio_out(I2SAudioOut *out) {
+    this->audio_out_ = out;
+    out->set_parent(this);
+  }
+
+  i2s_pin_config_t get_pin_config() const {
+    return {
+        .mck_io_num = I2S_PIN_NO_CHANGE,
+        .bck_io_num = this->bclk_pin_,
+        .ws_io_num = this->lrclk_pin_,
+        .data_out_num = I2S_PIN_NO_CHANGE,
+        .data_in_num = I2S_PIN_NO_CHANGE,
+    };
+  }
+
+  void set_bclk_pin(uint8_t pin) { this->bclk_pin_ = pin; }
+  void set_lrclk_pin(uint8_t pin) { this->lrclk_pin_ = pin; }
+
+  void lock() { this->lock_.lock(); }
+  bool try_lock() { return this->lock_.try_lock(); }
+  void unlock() { this->lock_.unlock(); }
+
+  i2s_port_t get_port() const { return this->port_; }
+
+ protected:
+  Mutex lock_;
+
+  I2SAudioIn *audio_in_{nullptr};
+  I2SAudioOut *audio_out_{nullptr};
+
+  uint8_t bclk_pin_;
+  uint8_t lrclk_pin_;
+  i2s_port_t port_{};
+};
+
+}  // namespace i2s_audio
+}  // namespace esphome
+
+#endif  // USE_ESP32
diff --git a/esphome/components/i2s_audio/media_player.py b/esphome/components/i2s_audio/media_player/__init__.py
similarity index 68%
rename from esphome/components/i2s_audio/media_player.py
rename to esphome/components/i2s_audio/media_player/__init__.py
index 43a48a721e..4ccb9cfc0a 100644
--- a/esphome/components/i2s_audio/media_player.py
+++ b/esphome/components/i2s_audio/media_player/__init__.py
@@ -5,22 +5,25 @@ import esphome.config_validation as cv
 from esphome import pins
 
 from esphome.const import CONF_ID, CONF_MODE
-from esphome.core import CORE
+
+from .. import (
+    i2s_audio_ns,
+    I2SAudioComponent,
+    I2SAudioOut,
+    CONF_I2S_AUDIO_ID,
+    CONF_I2S_DOUT_PIN,
+)
 
 CODEOWNERS = ["@jesserockz"]
-DEPENDENCIES = ["esp32"]
-
-i2s_audio_ns = cg.esphome_ns.namespace("i2s_audio")
+DEPENDENCIES = ["i2s_audio"]
 
 I2SAudioMediaPlayer = i2s_audio_ns.class_(
-    "I2SAudioMediaPlayer", cg.Component, media_player.MediaPlayer
+    "I2SAudioMediaPlayer", cg.Component, media_player.MediaPlayer, I2SAudioOut
 )
 
 i2s_dac_mode_t = cg.global_ns.enum("i2s_dac_mode_t")
 
-CONF_I2S_DOUT_PIN = "i2s_dout_pin"
-CONF_I2S_BCLK_PIN = "i2s_bclk_pin"
-CONF_I2S_LRCLK_PIN = "i2s_lrclk_pin"
+
 CONF_MUTE_PIN = "mute_pin"
 CONF_AUDIO_ID = "audio_id"
 CONF_DAC_TYPE = "dac_type"
@@ -48,34 +51,26 @@ def validate_esp32_variant(config):
 CONFIG_SCHEMA = cv.All(
     cv.typed_schema(
         {
-            "internal": cv.Schema(
+            "internal": media_player.MEDIA_PLAYER_SCHEMA.extend(
                 {
                     cv.GenerateID(): cv.declare_id(I2SAudioMediaPlayer),
+                    cv.GenerateID(CONF_I2S_AUDIO_ID): cv.use_id(I2SAudioComponent),
                     cv.Required(CONF_MODE): cv.enum(INTERNAL_DAC_OPTIONS, lower=True),
                 }
-            )
-            .extend(media_player.MEDIA_PLAYER_SCHEMA)
-            .extend(cv.COMPONENT_SCHEMA),
-            "external": cv.Schema(
+            ).extend(cv.COMPONENT_SCHEMA),
+            "external": media_player.MEDIA_PLAYER_SCHEMA.extend(
                 {
                     cv.GenerateID(): cv.declare_id(I2SAudioMediaPlayer),
+                    cv.GenerateID(CONF_I2S_AUDIO_ID): cv.use_id(I2SAudioComponent),
                     cv.Required(
                         CONF_I2S_DOUT_PIN
                     ): pins.internal_gpio_output_pin_number,
-                    cv.Required(
-                        CONF_I2S_BCLK_PIN
-                    ): pins.internal_gpio_output_pin_number,
-                    cv.Required(
-                        CONF_I2S_LRCLK_PIN
-                    ): pins.internal_gpio_output_pin_number,
                     cv.Optional(CONF_MUTE_PIN): pins.gpio_output_pin_schema,
                     cv.Optional(CONF_MODE, default="mono"): cv.one_of(
                         *EXTERNAL_DAC_OPTIONS, lower=True
                     ),
                 }
-            )
-            .extend(media_player.MEDIA_PLAYER_SCHEMA)
-            .extend(cv.COMPONENT_SCHEMA),
+            ).extend(cv.COMPONENT_SCHEMA),
         },
         key=CONF_DAC_TYPE,
     ),
@@ -89,19 +84,19 @@ async def to_code(config):
     await cg.register_component(var, config)
     await media_player.register_media_player(var, config)
 
+    parent = await cg.get_variable(config[CONF_I2S_AUDIO_ID])
+    cg.add(parent.register_audio_out(var))
+
     if config[CONF_DAC_TYPE] == "internal":
         cg.add(var.set_internal_dac_mode(config[CONF_MODE]))
     else:
         cg.add(var.set_dout_pin(config[CONF_I2S_DOUT_PIN]))
-        cg.add(var.set_bclk_pin(config[CONF_I2S_BCLK_PIN]))
-        cg.add(var.set_lrclk_pin(config[CONF_I2S_LRCLK_PIN]))
         if CONF_MUTE_PIN in config:
             pin = await cg.gpio_pin_expression(config[CONF_MUTE_PIN])
             cg.add(var.set_mute_pin(pin))
         cg.add(var.set_external_dac_channels(2 if config[CONF_MODE] == "stereo" else 1))
 
-    if CORE.is_esp32:
-        cg.add_library("WiFiClientSecure", None)
-        cg.add_library("HTTPClient", None)
-        cg.add_library("esphome/ESP32-audioI2S", "2.0.6")
-        cg.add_build_flag("-DAUDIO_NO_SD_FS")
+    cg.add_library("WiFiClientSecure", None)
+    cg.add_library("HTTPClient", None)
+    cg.add_library("esphome/ESP32-audioI2S", "2.0.6")
+    cg.add_build_flag("-DAUDIO_NO_SD_FS")
diff --git a/esphome/components/i2s_audio/i2s_audio_media_player.cpp b/esphome/components/i2s_audio/media_player/i2s_audio_media_player.cpp
similarity index 72%
rename from esphome/components/i2s_audio/i2s_audio_media_player.cpp
rename to esphome/components/i2s_audio/media_player/i2s_audio_media_player.cpp
index 2b00a5ec26..64f83a5ea6 100644
--- a/esphome/components/i2s_audio/i2s_audio_media_player.cpp
+++ b/esphome/components/i2s_audio/media_player/i2s_audio_media_player.cpp
@@ -11,11 +11,19 @@ static const char *const TAG = "audio";
 
 void I2SAudioMediaPlayer::control(const media_player::MediaPlayerCall &call) {
   if (call.get_media_url().has_value()) {
-    if (this->audio_->isRunning())
-      this->audio_->stopSong();
-    this->high_freq_.start();
-    this->audio_->connecttohost(call.get_media_url().value().c_str());
-    this->state = media_player::MEDIA_PLAYER_STATE_PLAYING;
+    this->current_url_ = call.get_media_url();
+
+    if (this->state == media_player::MEDIA_PLAYER_STATE_PLAYING && this->audio_ != nullptr) {
+      if (this->audio_->isRunning()) {
+        this->audio_->stopSong();
+      }
+      this->audio_->connecttohost(this->current_url_.value().c_str());
+    } else {
+      this->start();
+    }
+  }
+  if (this->i2s_state_ != I2S_STATE_RUNNING) {
+    return;
   }
   if (call.get_volume().has_value()) {
     this->volume = call.get_volume().value();
@@ -35,7 +43,7 @@ void I2SAudioMediaPlayer::control(const media_player::MediaPlayerCall &call) {
         this->state = media_player::MEDIA_PLAYER_STATE_PAUSED;
         break;
       case media_player::MEDIA_PLAYER_COMMAND_STOP:
-        this->stop_();
+        this->stop();
         break;
       case media_player::MEDIA_PLAYER_COMMAND_MUTE:
         this->mute_();
@@ -94,22 +102,51 @@ void I2SAudioMediaPlayer::set_volume_(float volume, bool publish) {
     this->volume = volume;
 }
 
-void I2SAudioMediaPlayer::stop_() {
-  if (this->audio_->isRunning())
-    this->audio_->stopSong();
-  this->high_freq_.stop();
+void I2SAudioMediaPlayer::setup() {
+  ESP_LOGCONFIG(TAG, "Setting up Audio...");
   this->state = media_player::MEDIA_PLAYER_STATE_IDLE;
 }
 
-void I2SAudioMediaPlayer::setup() {
-  ESP_LOGCONFIG(TAG, "Setting up Audio...");
+void I2SAudioMediaPlayer::loop() {
+  switch (this->i2s_state_) {
+    case I2S_STATE_STARTING:
+      this->start_();
+      break;
+    case I2S_STATE_RUNNING:
+      this->play_();
+      break;
+    case I2S_STATE_STOPPING:
+      this->stop_();
+      break;
+    case I2S_STATE_STOPPED:
+      break;
+  }
+}
+
+void I2SAudioMediaPlayer::play_() {
+  this->audio_->loop();
+  if (this->state == media_player::MEDIA_PLAYER_STATE_PLAYING && !this->audio_->isRunning()) {
+    this->stop();
+  }
+}
+
+void I2SAudioMediaPlayer::start() { this->i2s_state_ = I2S_STATE_STARTING; }
+void I2SAudioMediaPlayer::start_() {
+  if (this->parent_->try_lock()) {
+    return;  // Waiting for another i2s to return lock
+  }
+
 #if SOC_I2S_SUPPORTS_DAC
   if (this->internal_dac_mode_ != I2S_DAC_CHANNEL_DISABLE) {
-    this->audio_ = make_unique<Audio>(true, this->internal_dac_mode_);
+    this->audio_ = make_unique<Audio>(true, this->internal_dac_mode_, this->parent_->get_port());
   } else {
 #endif
-    this->audio_ = make_unique<Audio>(false);
-    this->audio_->setPinout(this->bclk_pin_, this->lrclk_pin_, this->dout_pin_);
+    this->audio_ = make_unique<Audio>(false, I2S_DAC_CHANNEL_BOTH_EN, this->parent_->get_port());
+
+    i2s_pin_config_t pin_config = this->parent_->get_pin_config();
+    pin_config.data_out_num = this->dout_pin_;
+    i2s_set_pin(this->parent_->get_port(), &pin_config);
+
     this->audio_->forceMono(this->external_dac_channels_ == 1);
     if (this->mute_pin_ != nullptr) {
       this->mute_pin_->setup();
@@ -118,16 +155,30 @@ void I2SAudioMediaPlayer::setup() {
 #if SOC_I2S_SUPPORTS_DAC
   }
 #endif
-  this->state = media_player::MEDIA_PLAYER_STATE_IDLE;
-}
-
-void I2SAudioMediaPlayer::loop() {
-  this->audio_->loop();
-  if (this->state == media_player::MEDIA_PLAYER_STATE_PLAYING && !this->audio_->isRunning()) {
-    this->stop_();
+  this->i2s_state_ = I2S_STATE_RUNNING;
+  this->high_freq_.start();
+  if (this->current_url_.has_value()) {
+    this->audio_->connecttohost(this->current_url_.value().c_str());
+    this->state = media_player::MEDIA_PLAYER_STATE_PLAYING;
     this->publish_state();
   }
 }
+void I2SAudioMediaPlayer::stop() { this->i2s_state_ = I2S_STATE_STOPPING; }
+void I2SAudioMediaPlayer::stop_() {
+  if (this->audio_->isRunning()) {
+    this->audio_->stopSong();
+    return;
+  }
+
+  this->audio_ = nullptr;
+  this->current_url_ = {};
+  this->parent_->unlock();
+  this->i2s_state_ = I2S_STATE_STOPPED;
+
+  this->high_freq_.stop();
+  this->state = media_player::MEDIA_PLAYER_STATE_IDLE;
+  this->publish_state();
+}
 
 media_player::MediaPlayerTraits I2SAudioMediaPlayer::get_traits() {
   auto traits = media_player::MediaPlayerTraits();
diff --git a/esphome/components/i2s_audio/i2s_audio_media_player.h b/esphome/components/i2s_audio/media_player/i2s_audio_media_player.h
similarity index 81%
rename from esphome/components/i2s_audio/i2s_audio_media_player.h
rename to esphome/components/i2s_audio/media_player/i2s_audio_media_player.h
index 40906d9ea5..dab9a85d7c 100644
--- a/esphome/components/i2s_audio/i2s_audio_media_player.h
+++ b/esphome/components/i2s_audio/media_player/i2s_audio_media_player.h
@@ -2,6 +2,10 @@
 
 #ifdef USE_ESP32_FRAMEWORK_ARDUINO
 
+#include "../i2s_audio.h"
+
+#include <driver/i2s.h>
+
 #include "esphome/components/media_player/media_player.h"
 #include "esphome/core/component.h"
 #include "esphome/core/gpio.h"
@@ -12,7 +16,14 @@
 namespace esphome {
 namespace i2s_audio {
 
-class I2SAudioMediaPlayer : public Component, public media_player::MediaPlayer {
+enum I2SState : uint8_t {
+  I2S_STATE_STOPPED = 0,
+  I2S_STATE_STARTING,
+  I2S_STATE_RUNNING,
+  I2S_STATE_STOPPING,
+};
+
+class I2SAudioMediaPlayer : public Component, public media_player::MediaPlayer, public I2SAudioOut {
  public:
   void setup() override;
   float get_setup_priority() const override { return esphome::setup_priority::LATE; }
@@ -22,8 +33,6 @@ class I2SAudioMediaPlayer : public Component, public media_player::MediaPlayer {
   void dump_config() override;
 
   void set_dout_pin(uint8_t pin) { this->dout_pin_ = pin; }
-  void set_bclk_pin(uint8_t pin) { this->bclk_pin_ = pin; }
-  void set_lrclk_pin(uint8_t pin) { this->lrclk_pin_ = pin; }
   void set_mute_pin(GPIOPin *mute_pin) { this->mute_pin_ = mute_pin; }
 #if SOC_I2S_SUPPORTS_DAC
   void set_internal_dac_mode(i2s_dac_mode_t mode) { this->internal_dac_mode_ = mode; }
@@ -34,20 +43,24 @@ class I2SAudioMediaPlayer : public Component, public media_player::MediaPlayer {
 
   bool is_muted() const override { return this->muted_; }
 
+  void start();
+  void stop();
+
  protected:
   void control(const media_player::MediaPlayerCall &call) override;
 
   void mute_();
   void unmute_();
   void set_volume_(float volume, bool publish = true);
-  void stop_();
 
+  void start_();
+  void stop_();
+  void play_();
+
+  I2SState i2s_state_{I2S_STATE_STOPPED};
   std::unique_ptr<Audio> audio_;
 
   uint8_t dout_pin_{0};
-  uint8_t din_pin_{0};
-  uint8_t bclk_pin_;
-  uint8_t lrclk_pin_;
 
   GPIOPin *mute_pin_{nullptr};
   bool muted_{false};
@@ -59,6 +72,8 @@ class I2SAudioMediaPlayer : public Component, public media_player::MediaPlayer {
   uint8_t external_dac_channels_;
 
   HighFrequencyLoopRequester high_freq_;
+
+  optional<std::string> current_url_{};
 };
 
 }  // namespace i2s_audio
diff --git a/esphome/components/i2s_audio/microphone/__init__.py b/esphome/components/i2s_audio/microphone/__init__.py
new file mode 100644
index 0000000000..08d7a5c95f
--- /dev/null
+++ b/esphome/components/i2s_audio/microphone/__init__.py
@@ -0,0 +1,41 @@
+import esphome.config_validation as cv
+import esphome.codegen as cg
+
+from esphome import pins
+from esphome.const import CONF_ID
+from esphome.components import microphone
+
+from .. import (
+    i2s_audio_ns,
+    I2SAudioComponent,
+    I2SAudioIn,
+    CONF_I2S_AUDIO_ID,
+    CONF_I2S_DIN_PIN,
+)
+
+CODEOWNERS = ["@jesserockz"]
+DEPENDENCIES = ["i2s_audio"]
+
+I2SAudioMicrophone = i2s_audio_ns.class_(
+    "I2SAudioMicrophone", I2SAudioIn, microphone.Microphone, cg.Component
+)
+
+CONFIG_SCHEMA = microphone.MICROPHONE_SCHEMA.extend(
+    {
+        cv.GenerateID(): cv.declare_id(I2SAudioMicrophone),
+        cv.GenerateID(CONF_I2S_AUDIO_ID): cv.use_id(I2SAudioComponent),
+        cv.Required(CONF_I2S_DIN_PIN): pins.internal_gpio_output_pin_number,
+    }
+).extend(cv.COMPONENT_SCHEMA)
+
+
+async def to_code(config):
+    var = cg.new_Pvariable(config[CONF_ID])
+    await cg.register_component(var, config)
+
+    parent = await cg.get_variable(config[CONF_I2S_AUDIO_ID])
+    cg.add(parent.register_audio_in(var))
+
+    cg.add(var.set_din_pin(config[CONF_I2S_DIN_PIN]))
+
+    await microphone.register_microphone(var, config)
diff --git a/esphome/components/i2s_audio/microphone/i2s_audio_microphone.cpp b/esphome/components/i2s_audio/microphone/i2s_audio_microphone.cpp
new file mode 100644
index 0000000000..8883cdc665
--- /dev/null
+++ b/esphome/components/i2s_audio/microphone/i2s_audio_microphone.cpp
@@ -0,0 +1,101 @@
+#include "i2s_audio_microphone.h"
+
+#ifdef USE_ESP32
+
+#include <driver/i2s.h>
+
+#include "esphome/core/hal.h"
+#include "esphome/core/log.h"
+
+namespace esphome {
+namespace i2s_audio {
+
+static const size_t BUFFER_SIZE = 512;
+
+static const char *const TAG = "i2s_audio.microphone";
+
+void I2SAudioMicrophone::setup() {
+  ESP_LOGCONFIG(TAG, "Setting up I2S Audio Microphone...");
+  this->buffer_.resize(BUFFER_SIZE);
+}
+
+void I2SAudioMicrophone::start() { this->state_ = microphone::STATE_STARTING; }
+void I2SAudioMicrophone::start_() {
+  if (!this->parent_->try_lock()) {
+    return;  // Waiting for another i2s to return lock
+  }
+  i2s_driver_config_t config = {
+      .mode = (i2s_mode_t) (I2S_MODE_MASTER | I2S_MODE_RX | I2S_MODE_PDM),
+      .sample_rate = 16000,
+      .bits_per_sample = I2S_BITS_PER_SAMPLE_16BIT,
+      .channel_format = I2S_CHANNEL_FMT_ONLY_RIGHT,
+      .communication_format = I2S_COMM_FORMAT_STAND_I2S,
+      .intr_alloc_flags = ESP_INTR_FLAG_LEVEL1,
+      .dma_buf_count = 4,
+      .dma_buf_len = 256,
+      .use_apll = false,
+      .tx_desc_auto_clear = false,
+      .fixed_mclk = 0,
+      .mclk_multiple = I2S_MCLK_MULTIPLE_DEFAULT,
+      .bits_per_chan = I2S_BITS_PER_CHAN_DEFAULT,
+  };
+
+  i2s_driver_install(this->parent_->get_port(), &config, 0, nullptr);
+
+  i2s_pin_config_t pin_config = this->parent_->get_pin_config();
+  pin_config.data_in_num = this->din_pin_;
+
+  i2s_set_pin(this->parent_->get_port(), &pin_config);
+  this->state_ = microphone::STATE_RUNNING;
+  this->high_freq_.start();
+}
+
+void I2SAudioMicrophone::stop() {
+  if (this->state_ == microphone::STATE_STOPPED)
+    return;
+  this->state_ = microphone::STATE_STOPPING;
+}
+
+void I2SAudioMicrophone::stop_() {
+  i2s_stop(this->parent_->get_port());
+  i2s_driver_uninstall(this->parent_->get_port());
+  this->parent_->unlock();
+  this->state_ = microphone::STATE_STOPPED;
+  this->high_freq_.stop();
+}
+
+void I2SAudioMicrophone::read_() {
+  size_t bytes_read = 0;
+  esp_err_t err =
+      i2s_read(this->parent_->get_port(), this->buffer_.data(), BUFFER_SIZE, &bytes_read, (100 / portTICK_PERIOD_MS));
+  if (err != ESP_OK) {
+    ESP_LOGW(TAG, "Error reading from I2S microphone: %s", esp_err_to_name(err));
+    this->status_set_warning();
+    return;
+  }
+
+  this->status_clear_warning();
+
+  this->data_callbacks_.call(this->buffer_);
+}
+
+void I2SAudioMicrophone::loop() {
+  switch (this->state_) {
+    case microphone::STATE_STOPPED:
+      break;
+    case microphone::STATE_STARTING:
+      this->start_();
+      break;
+    case microphone::STATE_RUNNING:
+      this->read_();
+      break;
+    case microphone::STATE_STOPPING:
+      this->stop_();
+      break;
+  }
+}
+
+}  // namespace i2s_audio
+}  // namespace esphome
+
+#endif  // USE_ESP32
diff --git a/esphome/components/i2s_audio/microphone/i2s_audio_microphone.h b/esphome/components/i2s_audio/microphone/i2s_audio_microphone.h
new file mode 100644
index 0000000000..a36cb7340c
--- /dev/null
+++ b/esphome/components/i2s_audio/microphone/i2s_audio_microphone.h
@@ -0,0 +1,37 @@
+#pragma once
+
+#ifdef USE_ESP32
+
+#include "../i2s_audio.h"
+
+#include "esphome/components/microphone/microphone.h"
+#include "esphome/core/component.h"
+
+namespace esphome {
+namespace i2s_audio {
+
+class I2SAudioMicrophone : public I2SAudioIn, public microphone::Microphone, public Component {
+ public:
+  void setup() override;
+  void start() override;
+  void stop() override;
+
+  void loop() override;
+
+  void set_din_pin(uint8_t pin) { this->din_pin_ = pin; }
+
+ protected:
+  void start_();
+  void stop_();
+  void read_();
+
+  uint8_t din_pin_{0};
+  std::vector<uint8_t> buffer_;
+
+  HighFrequencyLoopRequester high_freq_;
+};
+
+}  // namespace i2s_audio
+}  // namespace esphome
+
+#endif  // USE_ESP32
diff --git a/esphome/components/improv_base/improv_base.cpp b/esphome/components/improv_base/improv_base.cpp
index c95b3a77ec..f18a1061fb 100644
--- a/esphome/components/improv_base/improv_base.cpp
+++ b/esphome/components/improv_base/improv_base.cpp
@@ -21,7 +21,7 @@ std::string ImprovBase::get_formatted_next_url_() {
   // Ip address
   pos = this->next_url_.find("{{ip_address}}");
   if (pos != std::string::npos) {
-    std::string ip = network::IPAddress(network::get_ip_address()).str();
+    std::string ip = network::get_ip_address().str();
     copy.replace(pos, 14, ip);
   }
 
diff --git a/esphome/components/microphone/__init__.py b/esphome/components/microphone/__init__.py
new file mode 100644
index 0000000000..ff1f7aa963
--- /dev/null
+++ b/esphome/components/microphone/__init__.py
@@ -0,0 +1,91 @@
+from esphome import automation
+import esphome.config_validation as cv
+import esphome.codegen as cg
+
+from esphome.automation import maybe_simple_id
+from esphome.const import CONF_ID, CONF_TRIGGER_ID
+from esphome.core import CORE
+from esphome.coroutine import coroutine_with_priority
+
+
+CODEOWNERS = ["@jesserockz"]
+
+IS_PLATFORM_COMPONENT = True
+
+CONF_ON_DATA = "on_data"
+
+microphone_ns = cg.esphome_ns.namespace("microphone")
+
+Microphone = microphone_ns.class_("Microphone")
+
+CaptureAction = microphone_ns.class_(
+    "CaptureAction", automation.Action, cg.Parented.template(Microphone)
+)
+StopCaptureAction = microphone_ns.class_(
+    "StopCaptureAction", automation.Action, cg.Parented.template(Microphone)
+)
+
+
+DataTrigger = microphone_ns.class_(
+    "DataTrigger",
+    automation.Trigger.template(cg.std_vector.template(cg.int16).operator("ref")),
+)
+
+IsCapturingCondition = microphone_ns.class_(
+    "IsCapturingCondition", automation.Condition
+)
+
+
+async def setup_microphone_core_(var, config):
+    for conf in config.get(CONF_ON_DATA, []):
+        trigger = cg.new_Pvariable(conf[CONF_TRIGGER_ID], var)
+        await automation.build_automation(
+            trigger,
+            [(cg.std_vector.template(cg.uint8).operator("ref").operator("const"), "x")],
+            conf,
+        )
+
+
+async def register_microphone(var, config):
+    if not CORE.has_id(config[CONF_ID]):
+        var = cg.Pvariable(config[CONF_ID], var)
+    await setup_microphone_core_(var, config)
+
+
+MICROPHONE_SCHEMA = cv.Schema(
+    {
+        cv.Optional(CONF_ON_DATA): automation.validate_automation(
+            {
+                cv.GenerateID(CONF_TRIGGER_ID): cv.declare_id(DataTrigger),
+            }
+        ),
+    }
+)
+
+
+MICROPHONE_ACTION_SCHEMA = maybe_simple_id({cv.GenerateID(): cv.use_id(Microphone)})
+
+
+async def media_player_action(config, action_id, template_arg, args):
+    var = cg.new_Pvariable(action_id, template_arg)
+    await cg.register_parented(var, config[CONF_ID])
+    return var
+
+
+automation.register_action(
+    "microphone.capture", CaptureAction, MICROPHONE_ACTION_SCHEMA
+)(media_player_action)
+
+automation.register_action(
+    "microphone.stop_capture", StopCaptureAction, MICROPHONE_ACTION_SCHEMA
+)(media_player_action)
+
+automation.register_condition(
+    "microphone.is_capturing", IsCapturingCondition, MICROPHONE_ACTION_SCHEMA
+)(media_player_action)
+
+
+@coroutine_with_priority(100.0)
+async def to_code(config):
+    cg.add_global(microphone_ns.using)
+    cg.add_define("USE_MICROPHONE")
diff --git a/esphome/components/microphone/automation.h b/esphome/components/microphone/automation.h
new file mode 100644
index 0000000000..5f404b8d74
--- /dev/null
+++ b/esphome/components/microphone/automation.h
@@ -0,0 +1,32 @@
+#pragma once
+
+#include "esphome/core/automation.h"
+#include "microphone.h"
+
+#include <vector>
+
+namespace esphome {
+namespace microphone {
+
+template<typename... Ts> class CaptureAction : public Action<Ts...>, public Parented<Microphone> {
+  void play(Ts... x) override { this->parent_->start(); }
+};
+
+template<typename... Ts> class StopCaptureAction : public Action<Ts...>, public Parented<Microphone> {
+  void play(Ts... x) override { this->parent_->stop(); }
+};
+
+class DataTrigger : public Trigger<const std::vector<uint8_t> &> {
+ public:
+  explicit DataTrigger(Microphone *mic) {
+    mic->add_data_callback([this](const std::vector<uint8_t> &data) { this->trigger(data); });
+  }
+};
+
+template<typename... Ts> class IsCapturingActon : public Condition<Ts...>, public Parented<Microphone> {
+ public:
+  bool check(Ts... x) override { return this->parent_->is_running(); }
+};
+
+}  // namespace microphone
+}  // namespace esphome
diff --git a/esphome/components/microphone/microphone.h b/esphome/components/microphone/microphone.h
new file mode 100644
index 0000000000..b725f66ad7
--- /dev/null
+++ b/esphome/components/microphone/microphone.h
@@ -0,0 +1,33 @@
+#pragma once
+
+#include "esphome/core/entity_base.h"
+#include "esphome/core/helpers.h"
+
+namespace esphome {
+namespace microphone {
+
+enum State : uint8_t {
+  STATE_STOPPED = 0,
+  STATE_STARTING,
+  STATE_RUNNING,
+  STATE_STOPPING,
+};
+
+class Microphone {
+ public:
+  virtual void start() = 0;
+  virtual void stop() = 0;
+  void add_data_callback(std::function<void(const std::vector<uint8_t> &)> &&data_callback) {
+    this->data_callbacks_.add(std::move(data_callback));
+  }
+
+  bool is_running() const { return this->state_ == STATE_RUNNING; }
+
+ protected:
+  State state_{STATE_STOPPED};
+
+  CallbackManager<void(const std::vector<uint8_t> &)> data_callbacks_{};
+};
+
+}  // namespace microphone
+}  // namespace esphome
diff --git a/esphome/components/nfc/ndef_message.cpp b/esphome/components/nfc/ndef_message.cpp
index d7d134aedb..461856d377 100644
--- a/esphome/components/nfc/ndef_message.cpp
+++ b/esphome/components/nfc/ndef_message.cpp
@@ -44,6 +44,11 @@ NdefMessage::NdefMessage(std::vector<uint8_t> &data) {
       index += id_length;
     }
 
+    if ((data.begin() + index > data.end()) || (data.begin() + index + payload_length > data.end())) {
+      ESP_LOGE(TAG, "Corrupt record encountered; NdefMessage constructor aborting");
+      break;
+    }
+
     std::vector<uint8_t> payload_data(data.begin() + index, data.begin() + index + payload_length);
 
     std::unique_ptr<NdefRecord> record;
diff --git a/esphome/components/nfc/ndef_record.h b/esphome/components/nfc/ndef_record.h
index 27083fdaca..20542bf24b 100644
--- a/esphome/components/nfc/ndef_record.h
+++ b/esphome/components/nfc/ndef_record.h
@@ -42,8 +42,8 @@ class NdefRecord {
   virtual const std::string &get_payload() const { return this->payload_; };
 
   virtual std::vector<uint8_t> get_encoded_payload() {
-    std::vector<uint8_t> empty_payload;
-    return empty_payload;
+    std::vector<uint8_t> payload(this->payload_.begin(), this->payload_.end());
+    return payload;
   };
 
  protected:
diff --git a/esphome/components/nfc/nfc.cpp b/esphome/components/nfc/nfc.cpp
index 09dbdcfe94..b7c7215028 100644
--- a/esphome/components/nfc/nfc.cpp
+++ b/esphome/components/nfc/nfc.cpp
@@ -89,18 +89,18 @@ uint32_t get_mifare_classic_buffer_size(uint32_t message_length) {
 }
 
 bool mifare_classic_is_first_block(uint8_t block_num) {
-  if (block_num < 128) {
-    return (block_num % 4 == 0);
+  if (block_num < MIFARE_CLASSIC_BLOCKS_PER_SECT_LOW * MIFARE_CLASSIC_16BLOCK_SECT_START) {
+    return (block_num % MIFARE_CLASSIC_BLOCKS_PER_SECT_LOW == 0);
   } else {
-    return (block_num % 16 == 0);
+    return (block_num % MIFARE_CLASSIC_BLOCKS_PER_SECT_HIGH == 0);
   }
 }
 
 bool mifare_classic_is_trailer_block(uint8_t block_num) {
-  if (block_num < 128) {
-    return ((block_num + 1) % 4 == 0);
+  if (block_num < MIFARE_CLASSIC_BLOCKS_PER_SECT_LOW * MIFARE_CLASSIC_16BLOCK_SECT_START) {
+    return ((block_num + 1) % MIFARE_CLASSIC_BLOCKS_PER_SECT_LOW == 0);
   } else {
-    return ((block_num + 1) % 16 == 0);
+    return ((block_num + 1) % MIFARE_CLASSIC_BLOCKS_PER_SECT_HIGH == 0);
   }
 }
 
diff --git a/esphome/components/nfc/nfc.h b/esphome/components/nfc/nfc.h
index 1a4e8f1c1d..d4d66f970f 100644
--- a/esphome/components/nfc/nfc.h
+++ b/esphome/components/nfc/nfc.h
@@ -14,6 +14,9 @@ namespace nfc {
 static const uint8_t MIFARE_CLASSIC_BLOCK_SIZE = 16;
 static const uint8_t MIFARE_CLASSIC_LONG_TLV_SIZE = 4;
 static const uint8_t MIFARE_CLASSIC_SHORT_TLV_SIZE = 2;
+static const uint8_t MIFARE_CLASSIC_BLOCKS_PER_SECT_LOW = 4;
+static const uint8_t MIFARE_CLASSIC_BLOCKS_PER_SECT_HIGH = 16;
+static const uint8_t MIFARE_CLASSIC_16BLOCK_SECT_START = 32;
 
 static const uint8_t MIFARE_ULTRALIGHT_PAGE_SIZE = 4;
 static const uint8_t MIFARE_ULTRALIGHT_READ_SIZE = 4;
@@ -30,10 +33,18 @@ static const uint8_t TAG_TYPE_UNKNOWN = 99;
 // Mifare Commands
 static const uint8_t MIFARE_CMD_AUTH_A = 0x60;
 static const uint8_t MIFARE_CMD_AUTH_B = 0x61;
+static const uint8_t MIFARE_CMD_HALT = 0x50;
 static const uint8_t MIFARE_CMD_READ = 0x30;
 static const uint8_t MIFARE_CMD_WRITE = 0xA0;
 static const uint8_t MIFARE_CMD_WRITE_ULTRALIGHT = 0xA2;
 
+// Mifare Ack/Nak
+static const uint8_t MIFARE_CMD_ACK = 0x0A;
+static const uint8_t MIFARE_CMD_NAK_INVALID_XFER_BUFF_VALID = 0x00;
+static const uint8_t MIFARE_CMD_NAK_CRC_ERROR_XFER_BUFF_VALID = 0x01;
+static const uint8_t MIFARE_CMD_NAK_INVALID_XFER_BUFF_INVALID = 0x04;
+static const uint8_t MIFARE_CMD_NAK_CRC_ERROR_XFER_BUFF_INVALID = 0x05;
+
 static const char *const MIFARE_CLASSIC = "Mifare Classic";
 static const char *const NFC_FORUM_TYPE_2 = "NFC Forum Type 2";
 static const char *const ERROR = "Error";
diff --git a/esphome/components/pn532/pn532_mifare_classic.cpp b/esphome/components/pn532/pn532_mifare_classic.cpp
index 17d7aa123c..943f8c5519 100644
--- a/esphome/components/pn532/pn532_mifare_classic.cpp
+++ b/esphome/components/pn532/pn532_mifare_classic.cpp
@@ -52,7 +52,13 @@ std::unique_ptr<nfc::NfcTag> PN532::read_mifare_classic_tag_(std::vector<uint8_t
       current_block++;
     }
   }
-  buffer.erase(buffer.begin(), buffer.begin() + message_start_index);
+
+  if (buffer.begin() + message_start_index < buffer.end()) {
+    buffer.erase(buffer.begin(), buffer.begin() + message_start_index);
+  } else {
+    return make_unique<nfc::NfcTag>(uid, nfc::MIFARE_CLASSIC);
+  }
+
   return make_unique<nfc::NfcTag>(uid, nfc::MIFARE_CLASSIC, buffer);
 }
 
diff --git a/esphome/components/socket/bsd_sockets_impl.cpp b/esphome/components/socket/bsd_sockets_impl.cpp
index 3a08e3ea52..2dea4af277 100644
--- a/esphome/components/socket/bsd_sockets_impl.cpp
+++ b/esphome/components/socket/bsd_sockets_impl.cpp
@@ -139,6 +139,11 @@ class BSDSocketImpl : public Socket {
     return ::writev(fd_, iov, iovcnt);
 #endif
   }
+
+  ssize_t sendto(const void *buf, size_t len, int flags, const struct sockaddr *to, socklen_t tolen) override {
+    return ::sendto(fd_, buf, len, flags, to, tolen);
+  }
+
   int setblocking(bool blocking) override {
     int fl = ::fcntl(fd_, F_GETFL, 0);
     if (blocking) {
diff --git a/esphome/components/socket/lwip_raw_tcp_impl.cpp b/esphome/components/socket/lwip_raw_tcp_impl.cpp
index f5bb57bb93..bd59b81caa 100644
--- a/esphome/components/socket/lwip_raw_tcp_impl.cpp
+++ b/esphome/components/socket/lwip_raw_tcp_impl.cpp
@@ -467,6 +467,10 @@ class LWIPRawImpl : public Socket {
     }
     return written;
   }
+  ssize_t sendto(const void *buf, size_t len, int flags, const struct sockaddr *to, socklen_t tolen) override {
+    // return ::sendto(fd_, buf, len, flags, to, tolen);
+    return 0;
+  }
   int setblocking(bool blocking) override {
     if (pcb_ == nullptr) {
       errno = ECONNRESET;
diff --git a/esphome/components/socket/socket.cpp b/esphome/components/socket/socket.cpp
index 3fe7491d57..d00ddaeae2 100644
--- a/esphome/components/socket/socket.cpp
+++ b/esphome/components/socket/socket.cpp
@@ -1,7 +1,8 @@
 #include "socket.h"
-#include "esphome/core/log.h"
-#include <cstring>
 #include <cerrno>
+#include <cstring>
+#include <string>
+#include "esphome/core/log.h"
 
 namespace esphome {
 namespace socket {
@@ -14,7 +15,7 @@ std::unique_ptr<Socket> socket_ip(int type, int protocol) {
 #endif
 }
 
-socklen_t set_sockaddr(struct sockaddr *addr, socklen_t addrlen, const char *ip_address, uint16_t port) {
+socklen_t set_sockaddr(struct sockaddr *addr, socklen_t addrlen, const std::string &ip_address, uint16_t port) {
 #if LWIP_IPV6
   if (addrlen < sizeof(sockaddr_in6)) {
     errno = EINVAL;
@@ -24,9 +25,14 @@ socklen_t set_sockaddr(struct sockaddr *addr, socklen_t addrlen, const char *ip_
   memset(server, 0, sizeof(sockaddr_in6));
   server->sin6_family = AF_INET6;
   server->sin6_port = htons(port);
-  ip6_addr_t ip6;
-  inet6_aton(ip_address, &ip6);
-  memcpy(server->sin6_addr.un.u32_addr, ip6.addr, sizeof(ip6.addr));
+
+  if (ip_address.find('.') != std::string::npos) {
+    server->sin6_addr.un.u32_addr[3] = inet_addr(ip_address.c_str());
+  } else {
+    ip6_addr_t ip6;
+    inet6_aton(ip_address.c_str(), &ip6);
+    memcpy(server->sin6_addr.un.u32_addr, ip6.addr, sizeof(ip6.addr));
+  }
   return sizeof(sockaddr_in6);
 #else
   if (addrlen < sizeof(sockaddr_in)) {
@@ -36,7 +42,7 @@ socklen_t set_sockaddr(struct sockaddr *addr, socklen_t addrlen, const char *ip_
   auto *server = reinterpret_cast<sockaddr_in *>(addr);
   memset(server, 0, sizeof(sockaddr_in));
   server->sin_family = AF_INET;
-  server->sin_addr.s_addr = inet_addr(ip_address);
+  server->sin_addr.s_addr = inet_addr(ip_address.c_str());
   server->sin_port = htons(port);
   return sizeof(sockaddr_in);
 #endif
diff --git a/esphome/components/socket/socket.h b/esphome/components/socket/socket.h
index 175cb21a4a..7400ba306f 100644
--- a/esphome/components/socket/socket.h
+++ b/esphome/components/socket/socket.h
@@ -1,9 +1,9 @@
 #pragma once
-#include <string>
 #include <memory>
+#include <string>
 
-#include "headers.h"
 #include "esphome/core/optional.h"
+#include "headers.h"
 
 namespace esphome {
 namespace socket {
@@ -34,6 +34,8 @@ class Socket {
   virtual ssize_t readv(const struct iovec *iov, int iovcnt) = 0;
   virtual ssize_t write(const void *buf, size_t len) = 0;
   virtual ssize_t writev(const struct iovec *iov, int iovcnt) = 0;
+  virtual ssize_t sendto(const void *buf, size_t len, int flags, const struct sockaddr *to, socklen_t tolen);
+
   virtual int setblocking(bool blocking) = 0;
   virtual int loop() { return 0; };
 };
@@ -45,7 +47,7 @@ std::unique_ptr<Socket> socket(int domain, int type, int protocol);
 std::unique_ptr<Socket> socket_ip(int type, int protocol);
 
 /// Set a sockaddr to the specified address and port for the IP version used by socket_ip().
-socklen_t set_sockaddr(struct sockaddr *addr, socklen_t addrlen, const char *ip_address, uint16_t port);
+socklen_t set_sockaddr(struct sockaddr *addr, socklen_t addrlen, const std::string &ip_address, uint16_t port);
 
 /// Set a sockaddr to the any address and specified port for the IP version used by socket_ip().
 socklen_t set_sockaddr_any(struct sockaddr *addr, socklen_t addrlen, uint16_t port);
diff --git a/esphome/components/spi/spi.cpp b/esphome/components/spi/spi.cpp
index c3ffba4229..141bfb9448 100644
--- a/esphome/components/spi/spi.cpp
+++ b/esphome/components/spi/spi.cpp
@@ -79,7 +79,7 @@ void SPIComponent::setup() {
 #if defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32S2) || defined(USE_ESP32_VARIANT_ESP32S3)
       this->hw_spi_ = new SPIClass(FSPI);  // NOLINT(cppcoreguidelines-owning-memory)
 #else
-      this->hw_spi_ = new SPIClass(VSPI);  // NOLINT(cppcoreguidelines-owning-memory)
+      this->hw_spi_ = new SPIClass(HSPI);  // NOLINT(cppcoreguidelines-owning-memory)
 #endif  // USE_ESP32_VARIANT
     }
     spi_bus_num++;
diff --git a/esphome/components/ssd1306_base/__init__.py b/esphome/components/ssd1306_base/__init__.py
index f2e4ef5811..48143b9e1a 100644
--- a/esphome/components/ssd1306_base/__init__.py
+++ b/esphome/components/ssd1306_base/__init__.py
@@ -27,6 +27,7 @@ MODELS = {
     "SSD1306_96X16": SSD1306Model.SSD1306_MODEL_96_16,
     "SSD1306_64X48": SSD1306Model.SSD1306_MODEL_64_48,
     "SSD1306_64X32": SSD1306Model.SSD1306_MODEL_64_32,
+    "SSD1306_72X40": SSD1306Model.SSD1306_MODEL_72_40,
     "SH1106_128X32": SSD1306Model.SH1106_MODEL_128_32,
     "SH1106_128X64": SSD1306Model.SH1106_MODEL_128_64,
     "SH1106_96X16": SSD1306Model.SH1106_MODEL_96_16,
diff --git a/esphome/components/ssd1306_base/ssd1306_base.cpp b/esphome/components/ssd1306_base/ssd1306_base.cpp
index 2ba990637e..730e1c8f35 100644
--- a/esphome/components/ssd1306_base/ssd1306_base.cpp
+++ b/esphome/components/ssd1306_base/ssd1306_base.cpp
@@ -59,6 +59,7 @@ void SSD1306::setup() {
   // Set Y offset (0xD3)
   this->command(SSD1306_COMMAND_SET_DISPLAY_OFFSET_Y);
   this->command(0x00 + this->offset_y_);
+
   // Set start line at line 0 (0x40)
   this->command(SSD1306_COMMAND_SET_START_LINE | 0x00);
 
@@ -100,6 +101,7 @@ void SSD1306::setup() {
     case SH1107_MODEL_128_64:
     case SSD1305_MODEL_128_32:
     case SSD1305_MODEL_128_64:
+    case SSD1306_MODEL_72_40:
       this->command(0x12);
       break;
   }
@@ -118,6 +120,9 @@ void SSD1306::setup() {
     case SH1107_MODEL_128_64:
       this->command(0x35);
       break;
+    case SSD1306_MODEL_72_40:
+      this->command(0x20);
+      break;
     default:
       this->command(0x00);
       break;
@@ -156,6 +161,10 @@ void SSD1306::display() {
       this->command(0x20 + this->offset_x_);
       this->command(0x20 + this->offset_x_ + this->get_width_internal() - 1);
       break;
+    case SSD1306_MODEL_72_40:
+      this->command(0x1C + this->offset_x_);
+      this->command(0x1C + this->offset_x_ + this->get_width_internal() - 1);
+      break;
     default:
       this->command(0 + this->offset_x_);  // Page start address, 0
       this->command(this->get_width_internal() + this->offset_x_ - 1);
@@ -225,6 +234,8 @@ int SSD1306::get_height_internal() {
     case SSD1306_MODEL_64_48:
     case SH1106_MODEL_64_48:
       return 48;
+    case SSD1306_MODEL_72_40:
+      return 40;
     default:
       return 0;
   }
@@ -246,6 +257,8 @@ int SSD1306::get_width_internal() {
     case SH1106_MODEL_64_48:
     case SH1107_MODEL_128_64:
       return 64;
+    case SSD1306_MODEL_72_40:
+      return 72;
     default:
       return 0;
   }
@@ -294,6 +307,8 @@ const char *SSD1306::model_str_() {
       return "SSD1306 96x16";
     case SSD1306_MODEL_64_48:
       return "SSD1306 64x48";
+    case SSD1306_MODEL_72_40:
+      return "SSD1306 72x40";
     case SH1106_MODEL_128_32:
       return "SH1106 128x32";
     case SH1106_MODEL_128_64:
diff --git a/esphome/components/ssd1306_base/ssd1306_base.h b/esphome/components/ssd1306_base/ssd1306_base.h
index 2f5baffb69..7402ae3af2 100644
--- a/esphome/components/ssd1306_base/ssd1306_base.h
+++ b/esphome/components/ssd1306_base/ssd1306_base.h
@@ -13,6 +13,7 @@ enum SSD1306Model {
   SSD1306_MODEL_96_16,
   SSD1306_MODEL_64_48,
   SSD1306_MODEL_64_32,
+  SSD1306_MODEL_72_40,
   SH1106_MODEL_128_32,
   SH1106_MODEL_128_64,
   SH1106_MODEL_96_16,
diff --git a/esphome/components/ssd1306_i2c/ssd1306_i2c.cpp b/esphome/components/ssd1306_i2c/ssd1306_i2c.cpp
index 64b09c0672..96734eb618 100644
--- a/esphome/components/ssd1306_i2c/ssd1306_i2c.cpp
+++ b/esphome/components/ssd1306_i2c/ssd1306_i2c.cpp
@@ -53,8 +53,14 @@ void HOT I2CSSD1306::write_display_data() {
       }
     }
   } else {
+    size_t block_size = 16;
+    if ((this->get_buffer_length_() & 8) == 8) {
+      // use smaller block size for e.g. 72x40 displays where buffer size is multiple of 8, not 16
+      block_size = 8;
+    }
+
     for (uint32_t i = 0; i < this->get_buffer_length_();) {
-      uint8_t data[16];
+      uint8_t data[block_size];
       for (uint8_t &j : data)
         j = this->buffer_[i++];
       this->write_bytes(0x40, data, sizeof(data));
diff --git a/esphome/components/voice_assistant/__init__.py b/esphome/components/voice_assistant/__init__.py
new file mode 100644
index 0000000000..c90fd554ae
--- /dev/null
+++ b/esphome/components/voice_assistant/__init__.py
@@ -0,0 +1,57 @@
+import esphome.config_validation as cv
+import esphome.codegen as cg
+
+from esphome.const import CONF_ID, CONF_MICROPHONE
+from esphome import automation
+from esphome.automation import register_action
+from esphome.components import microphone
+
+AUTO_LOAD = ["socket"]
+DEPENDENCIES = ["api", "microphone"]
+
+CODEOWNERS = ["@jesserockz"]
+
+voice_assistant_ns = cg.esphome_ns.namespace("voice_assistant")
+VoiceAssistant = voice_assistant_ns.class_("VoiceAssistant", cg.Component)
+
+StartAction = voice_assistant_ns.class_(
+    "StartAction", automation.Action, cg.Parented.template(VoiceAssistant)
+)
+StopAction = voice_assistant_ns.class_(
+    "StopAction", automation.Action, cg.Parented.template(VoiceAssistant)
+)
+
+
+CONFIG_SCHEMA = cv.Schema(
+    {
+        cv.GenerateID(): cv.declare_id(VoiceAssistant),
+        cv.GenerateID(CONF_MICROPHONE): cv.use_id(microphone.Microphone),
+    }
+).extend(cv.COMPONENT_SCHEMA)
+
+
+async def to_code(config):
+    var = cg.new_Pvariable(config[CONF_ID])
+    await cg.register_component(var, config)
+
+    mic = await cg.get_variable(config[CONF_MICROPHONE])
+    cg.add(var.set_microphone(mic))
+
+    cg.add_define("USE_VOICE_ASSISTANT")
+
+
+VOICE_ASSISTANT_ACTION_SCHEMA = cv.Schema({cv.GenerateID(): cv.use_id(VoiceAssistant)})
+
+
+@register_action("voice_assistant.start", StartAction, VOICE_ASSISTANT_ACTION_SCHEMA)
+async def voice_assistant_listen_to_code(config, action_id, template_arg, args):
+    var = cg.new_Pvariable(action_id, template_arg)
+    await cg.register_parented(var, config[CONF_ID])
+    return var
+
+
+@register_action("voice_assistant.stop", StopAction, VOICE_ASSISTANT_ACTION_SCHEMA)
+async def voice_assistant_stop_to_code(config, action_id, template_arg, args):
+    var = cg.new_Pvariable(action_id, template_arg)
+    await cg.register_parented(var, config[CONF_ID])
+    return var
diff --git a/esphome/components/voice_assistant/voice_assistant.cpp b/esphome/components/voice_assistant/voice_assistant.cpp
new file mode 100644
index 0000000000..7f4bbf9934
--- /dev/null
+++ b/esphome/components/voice_assistant/voice_assistant.cpp
@@ -0,0 +1,148 @@
+#include "voice_assistant.h"
+
+#include "esphome/core/log.h"
+
+namespace esphome {
+namespace voice_assistant {
+
+static const char *const TAG = "voice_assistant";
+
+float VoiceAssistant::get_setup_priority() const { return setup_priority::AFTER_CONNECTION; }
+
+void VoiceAssistant::setup() {
+  ESP_LOGCONFIG(TAG, "Setting up Voice Assistant...");
+
+  global_voice_assistant = this;
+
+  this->socket_ = socket::socket(AF_INET, SOCK_DGRAM, IPPROTO_IP);
+  if (socket_ == nullptr) {
+    ESP_LOGW(TAG, "Could not create socket.");
+    this->mark_failed();
+    return;
+  }
+  int enable = 1;
+  int err = socket_->setsockopt(SOL_SOCKET, SO_REUSEADDR, &enable, sizeof(int));
+  if (err != 0) {
+    ESP_LOGW(TAG, "Socket unable to set reuseaddr: errno %d", err);
+    // we can still continue
+  }
+  err = socket_->setblocking(false);
+  if (err != 0) {
+    ESP_LOGW(TAG, "Socket unable to set nonblocking mode: errno %d", err);
+    this->mark_failed();
+    return;
+  }
+
+  this->mic_->add_data_callback([this](const std::vector<uint8_t> &data) {
+    if (!this->running_) {
+      return;
+    }
+    this->socket_->sendto(data.data(), data.size(), 0, (struct sockaddr *) &this->dest_addr_, sizeof(this->dest_addr_));
+  });
+}
+
+void VoiceAssistant::start(struct sockaddr_storage *addr, uint16_t port) {
+  ESP_LOGD(TAG, "Starting...");
+
+  memcpy(&this->dest_addr_, addr, sizeof(this->dest_addr_));
+  if (this->dest_addr_.ss_family == AF_INET) {
+    ((struct sockaddr_in *) &this->dest_addr_)->sin_port = htons(port);
+  }
+#if LWIP_IPV6
+  else if (this->dest_addr_.ss_family == AF_INET6) {
+    ((struct sockaddr_in6 *) &this->dest_addr_)->sin6_port = htons(port);
+  }
+#endif
+  else {
+    ESP_LOGW(TAG, "Unknown address family: %d", this->dest_addr_.ss_family);
+    return;
+  }
+  this->running_ = true;
+  this->mic_->start();
+}
+
+void VoiceAssistant::request_start() {
+  ESP_LOGD(TAG, "Requesting start...");
+  api::global_api_server->start_voice_assistant();
+}
+
+void VoiceAssistant::signal_stop() {
+  ESP_LOGD(TAG, "Signaling stop...");
+  this->mic_->stop();
+  this->running_ = false;
+  api::global_api_server->stop_voice_assistant();
+  memset(&this->dest_addr_, 0, sizeof(this->dest_addr_));
+}
+
+void VoiceAssistant::on_event(const api::VoiceAssistantEventResponse &msg) {
+  switch (msg.event_type) {
+    case api::enums::VOICE_ASSISTANT_RUN_END:
+      ESP_LOGD(TAG, "Voice Assistant ended.");
+      break;
+    case api::enums::VOICE_ASSISTANT_STT_END: {
+      std::string text;
+      for (auto arg : msg.data) {
+        if (arg.name == "text") {
+          text = std::move(arg.value);
+        }
+      }
+      if (text.empty()) {
+        ESP_LOGW(TAG, "No text in STT_END event.");
+        return;
+      }
+      ESP_LOGD(TAG, "Speech recognised as: \"%s\"", text.c_str());
+      // TODO `on_stt_end` trigger
+      break;
+    }
+    case api::enums::VOICE_ASSISTANT_TTS_START: {
+      std::string text;
+      for (auto arg : msg.data) {
+        if (arg.name == "text") {
+          text = std::move(arg.value);
+        }
+      }
+      if (text.empty()) {
+        ESP_LOGW(TAG, "No text in TTS_START event.");
+        return;
+      }
+      ESP_LOGD(TAG, "Response: \"%s\"", text.c_str());
+      // TODO `on_tts_start` trigger
+      break;
+    }
+    case api::enums::VOICE_ASSISTANT_TTS_END: {
+      std::string url;
+      for (auto arg : msg.data) {
+        if (arg.name == "url") {
+          url = std::move(arg.value);
+        }
+      }
+      if (url.empty()) {
+        ESP_LOGW(TAG, "No url in TTS_END event.");
+        return;
+      }
+      ESP_LOGD(TAG, "Response URL: \"%s\"", url.c_str());
+      // TODO `on_tts_end` trigger
+      break;
+    }
+    case api::enums::VOICE_ASSISTANT_ERROR: {
+      std::string code = "";
+      std::string message = "";
+      for (auto arg : msg.data) {
+        if (arg.name == "code") {
+          code = std::move(arg.value);
+        } else if (arg.name == "message") {
+          message = std::move(arg.value);
+        }
+      }
+      ESP_LOGE(TAG, "Error: %s - %s", code.c_str(), message.c_str());
+      // TODO `on_error` trigger
+    }
+    default:
+      break;
+  }
+}
+
+VoiceAssistant *global_voice_assistant = nullptr;  // NOLINT(cppcoreguidelines-avoid-non-const-global-variables)
+
+}  // namespace voice_assistant
+}  // namespace esphome
diff --git a/esphome/components/voice_assistant/voice_assistant.h b/esphome/components/voice_assistant/voice_assistant.h
new file mode 100644
index 0000000000..8cc20c31ad
--- /dev/null
+++ b/esphome/components/voice_assistant/voice_assistant.h
@@ -0,0 +1,50 @@
+#pragma once
+
+#include "esphome/core/automation.h"
+#include "esphome/core/component.h"
+#include "esphome/core/helpers.h"
+
+#include "esphome/components/api/api_pb2.h"
+#include "esphome/components/api/api_server.h"
+#include "esphome/components/microphone/microphone.h"
+#include "esphome/components/socket/socket.h"
+
+namespace esphome {
+namespace voice_assistant {
+
+class VoiceAssistant : public Component {
+ public:
+  void setup() override;
+  float get_setup_priority() const override;
+  void start(struct sockaddr_storage *addr, uint16_t port);
+
+  void set_microphone(microphone::Microphone *mic) { this->mic_ = mic; }
+
+  void request_start();
+  void signal_stop();
+
+  void on_event(const api::VoiceAssistantEventResponse &msg);
+
+ protected:
+  std::unique_ptr<socket::Socket> socket_ = nullptr;
+  struct sockaddr_storage dest_addr_;
+
+  microphone::Microphone *mic_{nullptr};
+
+  bool running_{false};
+};
+
+template<typename... Ts> class StartAction : public Action<Ts...>, public Parented<VoiceAssistant> {
+ public:
+  void play(Ts... x) override { this->parent_->request_start(); }
+};
+
+template<typename... Ts> class StopAction : public Action<Ts...>, public Parented<VoiceAssistant> {
+ public:
+  void play(Ts... x) override { this->parent_->signal_stop(); }
+};
+
+extern VoiceAssistant *global_voice_assistant;  // NOLINT(cppcoreguidelines-avoid-non-const-global-variables)
+
+}  // namespace voice_assistant
+}  // namespace esphome
diff --git a/esphome/const.py b/esphome/const.py
index f627b5abd4..534a8ade01 100644
--- a/esphome/const.py
+++ b/esphome/const.py
@@ -394,6 +394,7 @@ CONF_MEASUREMENT_SEQUENCE_NUMBER = "measurement_sequence_number"
 CONF_MEDIUM = "medium"
 CONF_MEMORY_BLOCKS = "memory_blocks"
 CONF_METHOD = "method"
+CONF_MICROPHONE = "microphone"
 CONF_MIN_COOLING_OFF_TIME = "min_cooling_off_time"
 CONF_MIN_COOLING_RUN_TIME = "min_cooling_run_time"
 CONF_MIN_FAN_MODE_SWITCHING_TIME = "min_fan_mode_switching_time"
diff --git a/esphome/core/defines.h b/esphome/core/defines.h
index c3738c0365..77d41e5b58 100644
--- a/esphome/core/defines.h
+++ b/esphome/core/defines.h
@@ -72,6 +72,7 @@
 #define USE_SOCKET_IMPL_BSD_SOCKETS
 #define USE_WIFI_11KV_SUPPORT
 #define USE_BLUETOOTH_PROXY
+#define USE_VOICE_ASSISTANT
 
 #ifdef USE_ARDUINO
 #define USE_ARDUINO_VERSION_CODE VERSION_CODE(2, 0, 5)
diff --git a/esphome/core/helpers.h b/esphome/core/helpers.h
index 8950cc3b8a..05a7eaa4cc 100644
--- a/esphome/core/helpers.h
+++ b/esphome/core/helpers.h
@@ -395,6 +395,9 @@ template<typename T, enable_if_t<std::is_unsigned<T>::value, int> = 0> std::stri
   val = convert_big_endian(val);
   return format_hex(reinterpret_cast<uint8_t *>(&val), sizeof(T));
 }
+template<std::size_t N> std::string format_hex(const std::array<uint8_t, N> &data) {
+  return format_hex(data.data(), data.size());
+}
 
 /// Format the byte array \p data of length \p len in pretty-printed, human-readable hex.
 std::string format_hex_pretty(const uint8_t *data, size_t length);
diff --git a/tests/test4.yaml b/tests/test4.yaml
index 54bb6f40a6..c21e71be00 100644
--- a/tests/test4.yaml
+++ b/tests/test4.yaml
@@ -653,13 +653,15 @@ touchscreen:
           format: Touch at (%d, %d)
           args: [touch.x, touch.y]
 
+i2s_audio:
+  i2s_lrclk_pin: GPIO26
+  i2s_bclk_pin: GPIO27
+
 media_player:
   - platform: i2s_audio
-    name: ${friendly_name}
+    name: None
     dac_type: external
-    i2s_lrclk_pin: GPIO26
     i2s_dout_pin: GPIO25
-    i2s_bclk_pin: GPIO27
     mute_pin: GPIO14
     on_state:
       - media_player.play:
@@ -685,3 +687,12 @@ prometheus:
     ha_hello_world:
       id: hellow_world
       name: Hello World
+
+microphone:
+  - platform: i2s_audio
+    id: mic_id
+    i2s_din_pin: GPIO23
+
+
+voice_assistant:
+  microphone: mic_id