diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index 1bfe5e64..74c0e525 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -107,6 +107,9 @@ target_sources(tomato PUBLIC
 	./frame_streams/audio_stream2.hpp
 	./frame_streams/stream_manager.hpp
 	./frame_streams/stream_manager.cpp
+	./frame_streams/locked_frame_stream.hpp
+
+	./frame_streams/voip_model.hpp
 
 	./frame_streams/sdl/sdl_audio2_frame_stream2.hpp
 	./frame_streams/sdl/sdl_audio2_frame_stream2.cpp
@@ -123,6 +126,9 @@ if (TOMATO_TOX_AV)
 	target_sources(tomato PUBLIC
 		./tox_av.hpp
 		./tox_av.cpp
+
+		./tox_av_voip_model.hpp
+		./tox_av_voip_model.cpp
 	)
 
 	target_compile_definitions(tomato PUBLIC TOMATO_TOX_AV)
diff --git a/src/chat_gui4.cpp b/src/chat_gui4.cpp
index b348ae42..ace44b76 100644
--- a/src/chat_gui4.cpp
+++ b/src/chat_gui4.cpp
@@ -9,6 +9,8 @@
 #include <solanaceae/contact/components.hpp>
 #include <solanaceae/util/utils.hpp>
 
+#include "./frame_streams/voip_model.hpp"
+
 // HACK: remove them
 #include <solanaceae/tox_contacts/components.hpp>
 
@@ -21,6 +23,7 @@
 
 #include "./media_meta_info_loader.hpp"
 #include "./sdl_clipboard_utils.hpp"
+#include "entt/entity/entity.hpp"
 
 #include <cctype>
 #include <ctime>
@@ -30,6 +33,7 @@
 #include <fstream>
 #include <iomanip>
 #include <sstream>
+#include <string>
 #include <variant>
 
 namespace Components {
@@ -257,6 +261,62 @@ float ChatGui4::render(float time_delta) {
 
 			if (ImGui::BeginChild(chat_label.c_str(), {0, 0}, ImGuiChildFlags_Border, ImGuiWindowFlags_MenuBar)) {
 				if (ImGui::BeginMenuBar()) {
+					// check if contact has voip model
+					// use activesessioncomp instead?
+					if (_cr.all_of<VoIPModelI*>(*_selected_contact)) {
+						if (ImGui::BeginMenu("VoIP")) {
+							auto* voip_model = _cr.get<VoIPModelI*>(*_selected_contact);
+
+							std::vector<Object> contact_sessions;
+							for (const auto& [ov, o_vm, sc] : _os.registry().view<VoIPModelI*, Components::VoIP::SessionContact>().each()) {
+								if (o_vm != voip_model) {
+									continue;
+								}
+								if (sc.c != *_selected_contact) {
+									continue;
+								}
+								contact_sessions.push_back(ov);
+							}
+
+							static VoIPModelI::DefaultConfig g_default_connections{};
+
+							if (ImGui::BeginMenu("default connections")) {
+								ImGui::MenuItem("incoming audio", nullptr, &g_default_connections.incoming_audio);
+								ImGui::MenuItem("incoming video", nullptr, &g_default_connections.incoming_video);
+								ImGui::Separator();
+								ImGui::MenuItem("outgoing audio", nullptr, &g_default_connections.outgoing_audio);
+								ImGui::MenuItem("outgoing video", nullptr, &g_default_connections.outgoing_video);
+								ImGui::EndMenu();
+							}
+
+							// TODO: only list if >1
+							if (ImGui::BeginMenu("accept call", false)) {
+								// list incomming here?
+								ImGui::EndMenu();
+							}
+
+							// TODO: disable if already in call?
+							if (ImGui::Button("call")) {
+								auto new_session = voip_model->enter(*_selected_contact, g_default_connections);
+							}
+
+							// TODO: only list if >1
+							if (ImGui::BeginMenu("leave/reject call", !contact_sessions.empty())) {
+								// list
+								for (const auto ov : contact_sessions) {
+									std::string label = "end #";
+									label += std::to_string(entt::to_integral(entt::to_entity(ov)));
+
+									if (ImGui::MenuItem(label.c_str())) {
+										voip_model->leave({_os.registry(), ov});
+									}
+								}
+								ImGui::EndMenu();
+							}
+
+							ImGui::EndMenu();
+						}
+					}
 					if (ImGui::BeginMenu("debug")) {
 						ImGui::Checkbox("show extra info", &_show_chat_extra_info);
 						ImGui::Checkbox("show avatar transfers", &_show_chat_avatar_tf);
diff --git a/src/frame_streams/locked_frame_stream.hpp b/src/frame_streams/locked_frame_stream.hpp
new file mode 100644
index 00000000..a2a43ebd
--- /dev/null
+++ b/src/frame_streams/locked_frame_stream.hpp
@@ -0,0 +1,42 @@
+#pragma once
+
+#include "./frame_stream2.hpp"
+
+#include <mutex>
+#include <deque>
+
+// threadsafe queue frame stream
+// protected by a simple mutex lock
+// prefer lockless queue implementations, when available
+template<typename FrameType>
+struct LockedFrameStream2 : public FrameStream2I<FrameType> {
+	std::mutex _lock;
+
+	std::deque<FrameType> _frames;
+
+	~LockedFrameStream2(void) {}
+
+	int32_t size(void) { return -1; }
+
+	std::optional<FrameType> pop(void) {
+		std::lock_guard lg{_lock};
+
+		if (_frames.empty()) {
+			return std::nullopt;
+		}
+
+		FrameType new_frame = std::move(_frames.front());
+		_frames.pop_front();
+
+		return std::move(new_frame);
+	}
+
+	bool push(const FrameType& value) {
+		std::lock_guard lg{_lock};
+
+		_frames.push_back(value);
+
+		return true;
+	}
+};
+
diff --git a/src/frame_streams/voip_model.hpp b/src/frame_streams/voip_model.hpp
index 6346aea8..83afeca5 100644
--- a/src/frame_streams/voip_model.hpp
+++ b/src/frame_streams/voip_model.hpp
@@ -10,9 +10,9 @@ namespace Components::VoIP {
 	struct TagVoIPSession {};
 
 	// to talk to the model handling this session
-	struct VoIPModel {
-		VoIPModelI* ptr {nullptr};
-	};
+	//struct VoIPModel {
+		//VoIPModelI* ptr {nullptr};
+	//};
 
 	struct SessionState {
 		// ????
@@ -59,13 +59,13 @@ struct VoIPModelI {
 		bool outgoing_audio {true};
 		bool outgoing_video {true};
 	};
-	virtual ObjectHandle enter(const Contact3 c, const DefaultConfig& defaults = {});
+	virtual ObjectHandle enter(const Contact3 c, const DefaultConfig& defaults = {true, true, true, true}) { return {}; }
 
 	// accept/join an invite to a session
-	virtual bool accept(ObjectHandle session);
+	virtual bool accept(ObjectHandle session) { return false; }
 
 	// leaves a call
 	// - VoIP session object
-	virtual bool leave(ObjectHandle session);
+	virtual bool leave(ObjectHandle session) { return false; }
 };
 
diff --git a/src/main_screen.cpp b/src/main_screen.cpp
index c6f70605..b552645a 100644
--- a/src/main_screen.cpp
+++ b/src/main_screen.cpp
@@ -25,13 +25,14 @@ MainScreen::MainScreen(SimpleConfigModel&& conf_, SDL_Renderer* renderer_, Theme
 	tc(save_path, save_password),
 	tpi(tc.getTox()),
 	ad(tc),
-#if TOMATO_TOX_AV
-	tav(tc.getTox()),
-#endif
 	tcm(cr, tc, tc),
 	tmm(rmm, cr, tcm, tc, tc),
 	ttm(rmm, cr, tcm, tc, tc, os),
 	tffom(cr, rmm, tcm, tc, tc),
+#if TOMATO_TOX_AV
+	tav(tc.getTox()),
+	tavvoip(os, tav, cr, tcm),
+#endif
 	theme(theme_),
 	mmil(rmm),
 	tam(/*rmm, */ os, cr, conf),
@@ -80,7 +81,7 @@ MainScreen::MainScreen(SimpleConfigModel&& conf_, SDL_Renderer* renderer_, Theme
 		g_provideInstance<ToxPrivateI>("ToxPrivateI", "host", &tpi);
 		g_provideInstance<ToxEventProviderI>("ToxEventProviderI", "host", &tc);
 #if TOMATO_TOX_AV
-		g_provideInstance<ToxAV>("ToxAV", "host", &tav);
+		g_provideInstance<ToxAVI>("ToxAVI", "host", &tav);
 #endif
 		g_provideInstance<ToxContactModel2>("ToxContactModel2", "host", &tcm);
 
@@ -528,6 +529,8 @@ Screen* MainScreen::tick(float time_delta, bool& quit) {
 	tav.toxavIterate();
 	// HACK: pow by 1.18 to increase 200 -> ~500
 	const float av_interval = std::pow(tav.toxavIterationInterval(), 1.18)/1000.f;
+
+	tavvoip.tick();
 #endif
 
 	tcm.iterate(time_delta); // compute
diff --git a/src/main_screen.hpp b/src/main_screen.hpp
index cd985a9a..daa3ab25 100644
--- a/src/main_screen.hpp
+++ b/src/main_screen.hpp
@@ -39,6 +39,7 @@
 
 #if TOMATO_TOX_AV
 #include "./tox_av.hpp"
+#include "./tox_av_voip_model.hpp"
 #endif
 
 #include <string>
@@ -67,13 +68,14 @@ struct MainScreen final : public Screen {
 	ToxClient tc;
 	ToxPrivateImpl tpi;
 	AutoDirty ad;
-#if TOMATO_TOX_AV
-	ToxAV tav;
-#endif
 	ToxContactModel2 tcm;
 	ToxMessageManager tmm;
 	ToxTransferManager ttm;
 	ToxFriendFauxOfflineMessaging tffom;
+#if TOMATO_TOX_AV
+	ToxAVI tav;
+	ToxAVVoIPModel tavvoip;
+#endif
 
 	Theme& theme;
 
diff --git a/src/tox_av.cpp b/src/tox_av.cpp
index 695744c3..4e6c044b 100644
--- a/src/tox_av.cpp
+++ b/src/tox_av.cpp
@@ -7,18 +7,7 @@
 
 // https://almogfx.bandcamp.com/track/crushed-w-cassade
 
-struct ToxAVFriendCallState final {
-	const uint32_t state {TOXAV_FRIEND_CALL_STATE_NONE};
-
-	[[nodiscard]] bool is_error(void) const { return state & TOXAV_FRIEND_CALL_STATE_ERROR; }
-	[[nodiscard]] bool is_finished(void) const { return state & TOXAV_FRIEND_CALL_STATE_FINISHED; }
-	[[nodiscard]] bool is_sending_a(void) const { return state & TOXAV_FRIEND_CALL_STATE_SENDING_A; }
-	[[nodiscard]] bool is_sending_v(void) const { return state & TOXAV_FRIEND_CALL_STATE_SENDING_V; }
-	[[nodiscard]] bool is_accepting_a(void) const { return state & TOXAV_FRIEND_CALL_STATE_ACCEPTING_A; }
-	[[nodiscard]] bool is_accepting_v(void) const { return state & TOXAV_FRIEND_CALL_STATE_ACCEPTING_V; }
-};
-
-ToxAV::ToxAV(Tox* tox) : _tox(tox) {
+ToxAVI::ToxAVI(Tox* tox) : _tox(tox) {
 	Toxav_Err_New err_new {TOXAV_ERR_NEW_OK};
 	_tox_av = toxav_new(_tox, &err_new);
 	// TODO: throw
@@ -28,7 +17,7 @@ ToxAV::ToxAV(Tox* tox) : _tox(tox) {
 		_tox_av,
 		+[](ToxAV*, uint32_t friend_number, bool audio_enabled, bool video_enabled, void *user_data) {
 			assert(user_data != nullptr);
-			static_cast<ToxAV*>(user_data)->cb_call(friend_number, audio_enabled, video_enabled);
+			static_cast<ToxAVI*>(user_data)->cb_call(friend_number, audio_enabled, video_enabled);
 		},
 		this
 	);
@@ -36,7 +25,7 @@ ToxAV::ToxAV(Tox* tox) : _tox(tox) {
 		_tox_av,
 		+[](ToxAV*, uint32_t friend_number, uint32_t state, void *user_data) {
 			assert(user_data != nullptr);
-			static_cast<ToxAV*>(user_data)->cb_call_state(friend_number, state);
+			static_cast<ToxAVI*>(user_data)->cb_call_state(friend_number, state);
 		},
 		this
 	);
@@ -44,7 +33,7 @@ ToxAV::ToxAV(Tox* tox) : _tox(tox) {
 		_tox_av,
 		+[](ToxAV*, uint32_t friend_number, uint32_t audio_bit_rate, void *user_data) {
 			assert(user_data != nullptr);
-			static_cast<ToxAV*>(user_data)->cb_audio_bit_rate(friend_number, audio_bit_rate);
+			static_cast<ToxAVI*>(user_data)->cb_audio_bit_rate(friend_number, audio_bit_rate);
 		},
 		this
 	);
@@ -52,7 +41,7 @@ ToxAV::ToxAV(Tox* tox) : _tox(tox) {
 		_tox_av,
 		+[](ToxAV*, uint32_t friend_number, uint32_t video_bit_rate, void *user_data) {
 			assert(user_data != nullptr);
-			static_cast<ToxAV*>(user_data)->cb_video_bit_rate(friend_number, video_bit_rate);
+			static_cast<ToxAVI*>(user_data)->cb_video_bit_rate(friend_number, video_bit_rate);
 		},
 		this
 	);
@@ -60,7 +49,7 @@ ToxAV::ToxAV(Tox* tox) : _tox(tox) {
 		_tox_av,
 		+[](ToxAV*, uint32_t friend_number, const int16_t pcm[], size_t sample_count, uint8_t channels, uint32_t sampling_rate, void *user_data) {
 			assert(user_data != nullptr);
-			static_cast<ToxAV*>(user_data)->cb_audio_receive_frame(friend_number, pcm, sample_count, channels, sampling_rate);
+			static_cast<ToxAVI*>(user_data)->cb_audio_receive_frame(friend_number, pcm, sample_count, channels, sampling_rate);
 		},
 		this
 	);
@@ -75,83 +64,83 @@ ToxAV::ToxAV(Tox* tox) : _tox(tox) {
 			void *user_data
 		) {
 			assert(user_data != nullptr);
-			static_cast<ToxAV*>(user_data)->cb_video_receive_frame(friend_number, width, height, y, u, v, ystride, ustride, vstride);
+			static_cast<ToxAVI*>(user_data)->cb_video_receive_frame(friend_number, width, height, y, u, v, ystride, ustride, vstride);
 		},
 		this
 	);
 }
 
-ToxAV::~ToxAV(void) {
+ToxAVI::~ToxAVI(void) {
 	toxav_kill(_tox_av);
 }
 
-uint32_t ToxAV::toxavIterationInterval(void) const {
+uint32_t ToxAVI::toxavIterationInterval(void) const {
 	return toxav_iteration_interval(_tox_av);
 }
 
-void ToxAV::toxavIterate(void) {
+void ToxAVI::toxavIterate(void) {
 	toxav_iterate(_tox_av);
 }
 
-uint32_t ToxAV::toxavAudioIterationInterval(void) const {
+uint32_t ToxAVI::toxavAudioIterationInterval(void) const {
 	return toxav_audio_iteration_interval(_tox_av);
 }
 
-void ToxAV::toxavAudioIterate(void) {
+void ToxAVI::toxavAudioIterate(void) {
 	toxav_audio_iterate(_tox_av);
 }
 
-uint32_t ToxAV::toxavVideoIterationInterval(void) const {
+uint32_t ToxAVI::toxavVideoIterationInterval(void) const {
 	return toxav_video_iteration_interval(_tox_av);
 }
 
-void ToxAV::toxavVideoIterate(void) {
+void ToxAVI::toxavVideoIterate(void) {
 	toxav_video_iterate(_tox_av);
 }
 
-Toxav_Err_Call ToxAV::toxavCall(uint32_t friend_number, uint32_t audio_bit_rate, uint32_t video_bit_rate) {
+Toxav_Err_Call ToxAVI::toxavCall(uint32_t friend_number, uint32_t audio_bit_rate, uint32_t video_bit_rate) {
 	Toxav_Err_Call err {TOXAV_ERR_CALL_OK};
 	toxav_call(_tox_av, friend_number, audio_bit_rate, video_bit_rate, &err);
 	return err;
 }
 
-Toxav_Err_Answer ToxAV::toxavAnswer(uint32_t friend_number, uint32_t audio_bit_rate, uint32_t video_bit_rate) {
+Toxav_Err_Answer ToxAVI::toxavAnswer(uint32_t friend_number, uint32_t audio_bit_rate, uint32_t video_bit_rate) {
 	Toxav_Err_Answer err {TOXAV_ERR_ANSWER_OK};
 	toxav_answer(_tox_av, friend_number, audio_bit_rate, video_bit_rate, &err);
 	return err;
 }
 
-Toxav_Err_Call_Control ToxAV::toxavCallControl(uint32_t friend_number, Toxav_Call_Control control) {
+Toxav_Err_Call_Control ToxAVI::toxavCallControl(uint32_t friend_number, Toxav_Call_Control control) {
 	Toxav_Err_Call_Control err {TOXAV_ERR_CALL_CONTROL_OK};
 	toxav_call_control(_tox_av, friend_number, control, &err);
 	return err;
 }
 
-Toxav_Err_Send_Frame ToxAV::toxavAudioSendFrame(uint32_t friend_number, const int16_t pcm[], size_t sample_count, uint8_t channels, uint32_t sampling_rate) {
+Toxav_Err_Send_Frame ToxAVI::toxavAudioSendFrame(uint32_t friend_number, const int16_t pcm[], size_t sample_count, uint8_t channels, uint32_t sampling_rate) {
 	Toxav_Err_Send_Frame err {TOXAV_ERR_SEND_FRAME_OK};
 	toxav_audio_send_frame(_tox_av, friend_number, pcm, sample_count, channels, sampling_rate, &err);
 	return err;
 }
 
-Toxav_Err_Bit_Rate_Set ToxAV::toxavAudioSetBitRate(uint32_t friend_number, uint32_t bit_rate) {
+Toxav_Err_Bit_Rate_Set ToxAVI::toxavAudioSetBitRate(uint32_t friend_number, uint32_t bit_rate) {
 	Toxav_Err_Bit_Rate_Set err {TOXAV_ERR_BIT_RATE_SET_OK};
 	toxav_audio_set_bit_rate(_tox_av, friend_number, bit_rate, &err);
 	return err;
 }
 
-Toxav_Err_Send_Frame ToxAV::toxavVideoSendFrame(uint32_t friend_number, uint16_t width, uint16_t height, const uint8_t y[], const uint8_t u[], const uint8_t v[]) {
+Toxav_Err_Send_Frame ToxAVI::toxavVideoSendFrame(uint32_t friend_number, uint16_t width, uint16_t height, const uint8_t y[], const uint8_t u[], const uint8_t v[]) {
 	Toxav_Err_Send_Frame err {TOXAV_ERR_SEND_FRAME_OK};
 	toxav_video_send_frame(_tox_av, friend_number, width, height, y, u, v, &err);
 	return err;
 }
 
-Toxav_Err_Bit_Rate_Set ToxAV::toxavVideoSetBitRate(uint32_t friend_number, uint32_t bit_rate) {
+Toxav_Err_Bit_Rate_Set ToxAVI::toxavVideoSetBitRate(uint32_t friend_number, uint32_t bit_rate) {
 	Toxav_Err_Bit_Rate_Set err {TOXAV_ERR_BIT_RATE_SET_OK};
 	toxav_video_set_bit_rate(_tox_av, friend_number, bit_rate, &err);
 	return err;
 }
 
-void ToxAV::cb_call(uint32_t friend_number, bool audio_enabled, bool video_enabled) {
+void ToxAVI::cb_call(uint32_t friend_number, bool audio_enabled, bool video_enabled) {
 	std::cerr << "TOXAV: receiving call f:" << friend_number << " a:" << audio_enabled << " v:" << video_enabled << "\n";
 	//Toxav_Err_Answer err_answer { TOXAV_ERR_ANSWER_OK };
 	//toxav_answer(_tox_av, friend_number, 0, 0, &err_answer);
@@ -169,7 +158,7 @@ void ToxAV::cb_call(uint32_t friend_number, bool audio_enabled, bool video_enabl
 	);
 }
 
-void ToxAV::cb_call_state(uint32_t friend_number, uint32_t state) {
+void ToxAVI::cb_call_state(uint32_t friend_number, uint32_t state) {
 	//ToxAVFriendCallState w_state{state};
 
 	//w_state.is_error();
@@ -185,7 +174,7 @@ void ToxAV::cb_call_state(uint32_t friend_number, uint32_t state) {
 	);
 }
 
-void ToxAV::cb_audio_bit_rate(uint32_t friend_number, uint32_t audio_bit_rate) {
+void ToxAVI::cb_audio_bit_rate(uint32_t friend_number, uint32_t audio_bit_rate) {
 	std::cerr << "TOXAV: audio bitrate f:" << friend_number << " abr:" << audio_bit_rate << "\n";
 
 	dispatch(
@@ -197,7 +186,7 @@ void ToxAV::cb_audio_bit_rate(uint32_t friend_number, uint32_t audio_bit_rate) {
 	);
 }
 
-void ToxAV::cb_video_bit_rate(uint32_t friend_number, uint32_t video_bit_rate) {
+void ToxAVI::cb_video_bit_rate(uint32_t friend_number, uint32_t video_bit_rate) {
 	std::cerr << "TOXAV: video bitrate f:" << friend_number << " vbr:" << video_bit_rate << "\n";
 
 	dispatch(
@@ -209,7 +198,7 @@ void ToxAV::cb_video_bit_rate(uint32_t friend_number, uint32_t video_bit_rate) {
 	);
 }
 
-void ToxAV::cb_audio_receive_frame(uint32_t friend_number, const int16_t pcm[], size_t sample_count, uint8_t channels, uint32_t sampling_rate) {
+void ToxAVI::cb_audio_receive_frame(uint32_t friend_number, const int16_t pcm[], size_t sample_count, uint8_t channels, uint32_t sampling_rate) {
 	//std::cerr << "TOXAV: audio frame f:" <<  friend_number << " sc:" << sample_count << " ch:" << (int)channels << " sr:" << sampling_rate << "\n";
 
 	dispatch(
@@ -223,7 +212,7 @@ void ToxAV::cb_audio_receive_frame(uint32_t friend_number, const int16_t pcm[],
 	);
 }
 
-void ToxAV::cb_video_receive_frame(
+void ToxAVI::cb_video_receive_frame(
 	uint32_t friend_number,
 	uint16_t width, uint16_t height,
 	const uint8_t y[/*! max(width, abs(ystride)) * height */],
diff --git a/src/tox_av.hpp b/src/tox_av.hpp
index bc00ce0f..7771118d 100644
--- a/src/tox_av.hpp
+++ b/src/tox_av.hpp
@@ -82,14 +82,15 @@ struct ToxAVEventI {
 };
 using ToxAVEventProviderI = EventProviderI<ToxAVEventI>;
 
-struct ToxAV : public ToxAVEventProviderI{
+// TODO: seperate out implementation from interface
+struct ToxAVI : public ToxAVEventProviderI {
 	Tox* _tox = nullptr;
 	ToxAV* _tox_av = nullptr;
 
 	static constexpr const char* version {"0"};
 
-	ToxAV(Tox* tox);
-	virtual ~ToxAV(void);
+	ToxAVI(Tox* tox);
+	virtual ~ToxAVI(void);
 
 	// interface
 	// if iterate is called on a different thread, it will fire events there
@@ -134,3 +135,14 @@ struct ToxAV : public ToxAVEventProviderI{
 	);
 };
 
+struct ToxAVFriendCallState final {
+	const uint32_t state {TOXAV_FRIEND_CALL_STATE_NONE};
+
+	[[nodiscard]] bool is_error(void) const { return state & TOXAV_FRIEND_CALL_STATE_ERROR; }
+	[[nodiscard]] bool is_finished(void) const { return state & TOXAV_FRIEND_CALL_STATE_FINISHED; }
+	[[nodiscard]] bool is_sending_a(void) const { return state & TOXAV_FRIEND_CALL_STATE_SENDING_A; }
+	[[nodiscard]] bool is_sending_v(void) const { return state & TOXAV_FRIEND_CALL_STATE_SENDING_V; }
+	[[nodiscard]] bool is_accepting_a(void) const { return state & TOXAV_FRIEND_CALL_STATE_ACCEPTING_A; }
+	[[nodiscard]] bool is_accepting_v(void) const { return state & TOXAV_FRIEND_CALL_STATE_ACCEPTING_V; }
+};
+
diff --git a/src/tox_av_voip_model.cpp b/src/tox_av_voip_model.cpp
new file mode 100644
index 00000000..f0f535dc
--- /dev/null
+++ b/src/tox_av_voip_model.cpp
@@ -0,0 +1,334 @@
+#include "./tox_av_voip_model.hpp"
+
+#include <solanaceae/object_store/object_store.hpp>
+#include <solanaceae/tox_contacts/components.hpp>
+
+#include "./frame_streams/stream_manager.hpp"
+#include "./frame_streams/audio_stream2.hpp"
+#include "./frame_streams/locked_frame_stream.hpp"
+
+#include <iostream>
+
+namespace Contact::Components {
+	// session instead???
+	struct ToxAVCall {
+		bool incoming_audio {false};
+		bool incoming_video {false};
+	};
+}
+
+namespace Components {
+	struct ToxAVAudioSink {
+		ObjectHandle o;
+		// ptr?
+	};
+	struct ToxAVAudioSource {
+		ObjectHandle o;
+		// ptr?
+	};
+} // Components
+
+struct ToxAVCallAudioSink : public FrameStream2SinkI<AudioFrame2> {
+	ToxAVI& _toxav;
+
+	// bitrate for enabled state
+	uint32_t _audio_bitrate {32};
+
+	uint32_t _fid;
+	std::shared_ptr<LockedFrameStream2<AudioFrame2>> _writer;
+
+	ToxAVCallAudioSink(ToxAVI& toxav, uint32_t fid) : _toxav(toxav), _fid(fid) {}
+	~ToxAVCallAudioSink(void) {
+		if (_writer) {
+			_writer = nullptr;
+			_toxav.toxavAudioSetBitRate(_fid, 0);
+		}
+	}
+
+	// sink
+	std::shared_ptr<FrameStream2I<AudioFrame2>> subscribe(void) override {
+		if (_writer) {
+			// max 1 (exclusive for now)
+			return nullptr;
+		}
+
+		auto err = _toxav.toxavAudioSetBitRate(_fid, _audio_bitrate);
+		if (err != TOXAV_ERR_BIT_RATE_SET_OK) {
+			return nullptr;
+		}
+
+		_writer = std::make_shared<LockedFrameStream2<AudioFrame2>>();
+
+		return _writer;
+	}
+
+	bool unsubscribe(const std::shared_ptr<FrameStream2I<AudioFrame2>>& sub) override {
+		if (!sub || !_writer) {
+			// nah
+			return false;
+		}
+
+		if (sub == _writer) {
+			_writer = nullptr;
+
+			/*auto err = */_toxav.toxavAudioSetBitRate(_fid, 0);
+			// print warning? on error?
+
+			return true;
+		}
+
+		// what
+		return false;
+	}
+};
+
+ToxAVVoIPModel::ToxAVVoIPModel(ObjectStore2& os, ToxAVI& av, Contact3Registry& cr, ToxContactModel2& tcm) :
+	_os(os), _av(av), _cr(cr), _tcm(tcm)
+{
+	_av.subscribe(this, ToxAV_Event::friend_call);
+	_av.subscribe(this, ToxAV_Event::friend_call_state);
+	_av.subscribe(this, ToxAV_Event::friend_audio_bitrate);
+	_av.subscribe(this, ToxAV_Event::friend_video_bitrate);
+	_av.subscribe(this, ToxAV_Event::friend_audio_frame);
+	_av.subscribe(this, ToxAV_Event::friend_video_frame);
+
+	// attach to all tox friend contacts
+
+	for (const auto& [cv, _] : _cr.view<Contact::Components::ToxFriendPersistent>().each()) {
+		_cr.emplace<VoIPModelI*>(cv, this);
+	}
+	// TODO: events
+}
+
+ToxAVVoIPModel::~ToxAVVoIPModel(void) {
+	for (const auto& [ov, voipmodel] : _os.registry().view<VoIPModelI*>().each()) {
+		if (voipmodel == this) {
+			destroySession(_os.objectHandle(ov));
+		}
+	}
+}
+
+void ToxAVVoIPModel::destroySession(ObjectHandle session) {
+	if (!static_cast<bool>(session)) {
+		return;
+	}
+
+	// destory sources
+	if (auto* ss = session.try_get<Components::VoIP::StreamSources>(); ss != nullptr) {
+		for (const auto ssov : ss->streams) {
+
+			_os.throwEventDestroy(ssov);
+			_os.registry().destroy(ssov);
+		}
+	}
+
+	// destory sinks
+	if (auto* ss = session.try_get<Components::VoIP::StreamSinks>(); ss != nullptr) {
+		for (const auto ssov : ss->streams) {
+
+			_os.throwEventDestroy(ssov);
+			_os.registry().destroy(ssov);
+		}
+	}
+
+	// destory session
+	_os.throwEventDestroy(session);
+	_os.registry().destroy(session);
+}
+
+void ToxAVVoIPModel::tick(void) {
+	//for (const auto& [oc, asink, asrf] : _os.registry().view<ToxAVCallAudioSink*, AudioStreamReFramer>().each()) {
+	for (const auto& [oc, asink] : _os.registry().view<ToxAVCallAudioSink*>().each()) {
+		if (!asink->_writer) {
+			continue;
+		}
+
+		//asrf._stream = asink->_writer.get();
+
+		for (size_t i = 0; i < 10; i++) {
+			auto new_frame_opt = asink->_writer->pop();
+			//auto new_frame_opt = asrf.pop();
+			if (!new_frame_opt.has_value()) {
+				break;
+			}
+			const auto& new_frame = new_frame_opt.value();
+
+			//* @param sample_count Number of samples in this frame. Valid numbers here are
+			//*   `((sample rate) * (audio length) / 1000)`, where audio length can be
+			//*   2.5, 5, 10, 20, 40 or 60 milliseconds.
+
+			// we likely needs to subdivide/repackage
+			// frame size should be an option exposed to the user
+			// with 10ms as a default ?
+			// the larger the frame size, the less overhead but the more delay
+
+			auto err = _av.toxavAudioSendFrame(
+				asink->_fid,
+				new_frame.getSpan().ptr,
+				new_frame.getSpan().size / new_frame.channels,
+				new_frame.channels,
+				new_frame.sample_rate
+			);
+			if (err != TOXAV_ERR_SEND_FRAME_OK) {
+				std::cerr << "DTC: failed to send audio frame " << err << "\n";
+			}
+		}
+	}
+}
+
+ObjectHandle ToxAVVoIPModel::enter(const Contact3 c, const DefaultConfig& defaults) {
+	if (!_cr.all_of<Contact::Components::ToxFriendEphemeral>(c)) {
+		return {};
+	}
+
+	const auto friend_number = _cr.get<Contact::Components::ToxFriendEphemeral>(c).friend_number;
+
+	std::cerr << "TAVVOIP: lol\n";
+	const auto err = _av.toxavCall(friend_number, 0, 0);
+	if (err != TOXAV_ERR_CALL_OK) {
+		std::cerr << "TAVVOIP error: failed to start call: " << err << "\n";
+		return {};
+	}
+
+	ObjectHandle new_session {_os.registry(), _os.registry().create()};
+
+	new_session.emplace<Components::VoIP::SessionContact>(c);
+	new_session.emplace<Components::VoIP::TagVoIPSession>(); // ??
+	new_session.emplace<Components::VoIP::SessionState>().state = Components::VoIP::SessionState::State::RINGING;
+	new_session.emplace<VoIPModelI*>(this);
+
+	_os.throwEventConstruct(new_session);
+	return new_session;
+}
+
+bool ToxAVVoIPModel::accept(ObjectHandle session) {
+	//_av.toxavAnswer(, 0, 0);
+	return false;
+}
+
+bool ToxAVVoIPModel::leave(ObjectHandle session) {
+	// rename to end?
+
+	if (!static_cast<bool>(session)) {
+		return false;
+	}
+
+	if (!session.all_of<Components::VoIP::TagVoIPSession, VoIPModelI*, Components::VoIP::SessionContact>()) {
+		return false;
+	}
+
+	// check if self
+	if (session.get<VoIPModelI*>() != this) {
+		return false;
+	}
+
+	const auto session_contact = session.get<Components::VoIP::SessionContact>().c;
+	if (!_cr.all_of<Contact::Components::ToxFriendEphemeral>(session_contact)) {
+		return false;
+	}
+
+	const auto friend_number = _cr.get<Contact::Components::ToxFriendEphemeral>(session_contact).friend_number;
+	// check error? (we delete anyway)
+	_av.toxavCallControl(friend_number, Toxav_Call_Control::TOXAV_CALL_CONTROL_CANCEL);
+
+	destroySession(session);
+	return true;
+}
+
+bool ToxAVVoIPModel::onEvent(const Events::FriendCall& e) {
+	// new incoming call, create voip session, ready to be accepted
+	// (or rejected...)
+
+	const auto session_contact = _tcm.getContactFriend(e.friend_number);
+	if (!_cr.valid(session_contact)) {
+		return false;
+	}
+
+	ObjectHandle new_session {_os.registry(), _os.registry().create()};
+
+	new_session.emplace<Components::VoIP::SessionContact>(session_contact);
+	new_session.emplace<Components::VoIP::TagVoIPSession>(); // ??
+	new_session.emplace<Components::VoIP::SessionState>().state = Components::VoIP::SessionState::State::RINGING;
+	new_session.emplace<VoIPModelI*>(this);
+
+	_os.throwEventConstruct(new_session);
+	return true;
+}
+
+bool ToxAVVoIPModel::onEvent(const Events::FriendCallState& e) {
+	const auto session_contact = _tcm.getContactFriend(e.friend_number);
+	if (!_cr.valid(session_contact)) {
+		return false;
+	}
+
+	ToxAVFriendCallState s{e.state};
+
+	// find session(s?)
+	// TODO: keep lookup table
+	for (const auto& [ov, voipmodel] : _os.registry().view<VoIPModelI*>().each()) {
+		if (voipmodel == this) {
+			auto o = _os.objectHandle(ov);
+
+			if (!o.all_of<Components::VoIP::SessionContact>()) {
+				continue;
+			}
+			if (session_contact != o.get<Components::VoIP::SessionContact>().c) {
+				continue;
+			}
+
+			if (s.is_error() || s.is_finished()) {
+				// destroy call
+				destroySession(o);
+			} else {
+				// remote accepted our call, or av send/recv conditions changed?
+				o.get<Components::VoIP::SessionState>().state; // set to in call
+
+				auto& stream_sinks = o.get_or_emplace<Components::VoIP::StreamSinks>().streams;
+				if (s.is_accepting_a() && !o.all_of<Components::ToxAVAudioSink>()) {
+					ObjectHandle outgoing_audio {_os.registry(), _os.registry().create()};
+
+					auto new_asink = std::make_unique<ToxAVCallAudioSink>(_av, e.friend_number);
+					outgoing_audio.emplace<ToxAVCallAudioSink*>(new_asink.get());
+					//outgoing_audio.emplace<AudioStreamReFramer>().frame_length_ms = 10;
+					outgoing_audio.emplace<Components::FrameStream2Sink<AudioFrame2>>(std::move(new_asink));
+					outgoing_audio.emplace<Components::StreamSink>(Components::StreamSink::create<AudioFrame2>("ToxAV Friend Call Outgoing Audio"));
+					outgoing_audio.emplace<Components::TagConnectToDefault>(); // depends on what was specified in enter()
+
+					stream_sinks.push_back(outgoing_audio);
+					o.emplace<Components::ToxAVAudioSink>(outgoing_audio);
+					// TODO: tie session to stream
+
+					_os.throwEventConstruct(outgoing_audio);
+				} else if (!s.is_accepting_a() && o.all_of<Components::ToxAVAudioSink>()) {
+					// remove asink?
+				}
+
+				// video
+
+				// add/update sources
+				// audio
+				// video
+
+			}
+		}
+	}
+
+	return true;
+}
+
+bool ToxAVVoIPModel::onEvent(const Events::FriendAudioBitrate&) {
+	return false;
+}
+
+bool ToxAVVoIPModel::onEvent(const Events::FriendVideoBitrate&) {
+	return false;
+}
+
+bool ToxAVVoIPModel::onEvent(const Events::FriendAudioFrame&) {
+	return false;
+}
+
+bool ToxAVVoIPModel::onEvent(const Events::FriendVideoFrame&) {
+	return false;
+}
+
diff --git a/src/tox_av_voip_model.hpp b/src/tox_av_voip_model.hpp
new file mode 100644
index 00000000..f067fc88
--- /dev/null
+++ b/src/tox_av_voip_model.hpp
@@ -0,0 +1,35 @@
+#pragma once
+
+#include <solanaceae/object_store/fwd.hpp>
+#include <solanaceae/contact/contact_model3.hpp>
+#include <solanaceae/tox_contacts/tox_contact_model2.hpp>
+#include "./frame_streams/voip_model.hpp"
+#include "./tox_av.hpp"
+
+struct ToxAVVoIPModel : protected ToxAVEventI, public VoIPModelI {
+	ObjectStore2& _os;
+	ToxAVI& _av;
+	Contact3Registry& _cr;
+	ToxContactModel2& _tcm;
+
+	ToxAVVoIPModel(ObjectStore2& os, ToxAVI& av, Contact3Registry& cr, ToxContactModel2& tcm);
+	~ToxAVVoIPModel(void);
+
+	void destroySession(ObjectHandle session);
+
+	void tick(void);
+
+	public: // voip model
+		ObjectHandle enter(const Contact3 c, const DefaultConfig& defaults) override;
+		bool accept(ObjectHandle session) override;
+		bool leave(ObjectHandle session) override;
+
+	protected: // toxav events
+		bool onEvent(const Events::FriendCall&) override;
+		bool onEvent(const Events::FriendCallState&) override;
+		bool onEvent(const Events::FriendAudioBitrate&) override;
+		bool onEvent(const Events::FriendVideoBitrate&) override;
+		bool onEvent(const Events::FriendAudioFrame&) override;
+		bool onEvent(const Events::FriendVideoFrame&) override;
+};
+