diff --git a/flywms_navigation_gui.py b/flywms_navigation_gui.py
new file mode 100644
index 0000000..871bace
--- /dev/null
+++ b/flywms_navigation_gui.py
@@ -0,0 +1,403 @@
+import time
+from dataclasses import dataclass
+from pathlib import Path
+
+import cv2
+import numpy as np
+
+import flywms_navigation as nav
+
+try:
+    import dearpygui.dearpygui as dpg
+except ModuleNotFoundError as exc:
+    dpg = None
+    DEARPYGUI_IMPORT_ERROR = exc
+else:
+    DEARPYGUI_IMPORT_ERROR = None
+
+
+TEXTURE_MAIN = "texture_main"
+TEXTURE_SNAPSHOT = "texture_snapshot"
+TEXTURE_REGISTRY = "texture_registry"
+IMAGE_MAIN = "image_main"
+IMAGE_SNAPSHOT = "image_snapshot"
+TEXT_STATUS = "text_status"
+TEXT_COMMANDS = "text_commands"
+TEXT_TRACKS = "text_tracks"
+TEXT_METRICS = "text_metrics"
+BUTTON_RUN = "button_run"
+
+
+@dataclass(frozen=True)
+class DemoFrameState:
+    frame_id: int
+    elapsed: float
+    source_frame: np.ndarray
+    display_frame: np.ndarray
+    snapshot_frame: np.ndarray | None
+    tracks: list[nav.Track]
+    detections_count: int
+    snapshots: list[nav.NavigationSnapshot]
+    yolo_ms: float
+    fps_text: str
+
+
+class NavigationDemoEngine:
+    """One-step navigation simulator used by the DearPyGUI shell.
+
+    The original flywms_navigation.py command-line program stays untouched.
+    This class reuses its detector/tracker/navigation objects and exposes a
+    frame-by-frame API suitable for a GUI event loop.
+    """
+
+    def __init__(self, args):
+        self.args = args
+        nav.require_file(args.weights, "modello Ultralytics")
+        self.detector = nav.UltralyticsDetector(args.weights, args.ultralytics_device)
+        self.cap, self.source_name = nav.open_capture(args.video)
+        if not self.cap.isOpened():
+            raise RuntimeError(f"impossibile aprire sorgente video: {self.source_name}")
+
+        video_fps = self.cap.get(cv2.CAP_PROP_FPS)
+        self.frame_delay = (
+            1.0 / video_fps
+            if args.realtime_playback and video_fps and video_fps > 1
+            else 0.0
+        )
+        self.tracker = nav.LightweightTracker(
+            max_missed=args.max_track_missed,
+            min_match_score=args.min_match_score,
+            max_center_distance_ratio=args.max_center_distance_ratio,
+        )
+        self.navigator = nav.NavigationController(args)
+        self.frame_id = 0
+        self.start_time = time.perf_counter()
+        self.last_loop_end = self.start_time
+        self.yolo_total_ms = 0.0
+        self.yolo_cycles = 0
+        self.stop_reason = ""
+
+    def close(self) -> None:
+        self.cap.release()
+
+    def step(self) -> DemoFrameState | None:
+        if self.frame_delay > 0:
+            now = time.perf_counter()
+            sleep_for = self.frame_delay - (now - self.last_loop_end)
+            if sleep_for > 0:
+                time.sleep(sleep_for)
+            self.last_loop_end = time.perf_counter()
+
+        ok, frame = self.cap.read()
+        if not ok:
+            self.stop_reason = "Fine stream"
+            return None
+
+        self.frame_id += 1
+        timestamp = time.perf_counter()
+        if self.args.max_frames > 0 and self.frame_id > self.args.max_frames:
+            self.stop_reason = f"Raggiunto max_frames={self.args.max_frames}"
+            return None
+
+        detections, yolo_ms = self.detector.detect(
+            frame,
+            self.args.min_confidence,
+            self.args.input_size,
+        )
+        self.yolo_total_ms += yolo_ms
+        self.yolo_cycles += 1
+
+        gaylords = [
+            det for det in detections
+            if det.class_name.strip().lower() == self.args.target_class.strip().lower()
+        ]
+        tracks = self.tracker.update(gaylords, self.frame_id, frame.shape[1])
+
+        if (
+            self.args.motion_report_interval > 0
+            and self.frame_id % self.args.motion_report_interval == 0
+        ):
+            self.navigator.set_motion_text(
+                nav.estimate_motion_from_tracks(tracks, self.args.motion_min_pixels)
+            )
+
+        new_snapshots: list[nav.NavigationSnapshot] = []
+        for track in tracks:
+            if track.missed == 0:
+                snapshot = self.navigator.process_track(
+                    track,
+                    frame,
+                    self.frame_id,
+                    timestamp,
+                )
+                if snapshot is not None:
+                    new_snapshots.append(snapshot)
+
+        for snapshot in new_snapshots:
+            self.navigator.simulate_remote_response(snapshot)
+
+        elapsed = max(time.perf_counter() - self.start_time, 0.001)
+        fps_text = (
+            f"frame={self.frame_id} fps={self.frame_id / elapsed:.1f} "
+            f"det={len(gaylords)} tracks={len(tracks)} "
+            f"snap={self.navigator.snapshot_counter}"
+        )
+        display = nav.draw_navigation_debug(
+            frame,
+            tracks,
+            self.args,
+            self.navigator.last_command_text,
+            fps_text,
+        )
+
+        return DemoFrameState(
+            frame_id=self.frame_id,
+            elapsed=elapsed,
+            source_frame=frame,
+            display_frame=display,
+            snapshot_frame=self.navigator.last_ocr_payload_frame,
+            tracks=tracks,
+            detections_count=len(gaylords),
+            snapshots=new_snapshots,
+            yolo_ms=yolo_ms,
+            fps_text=fps_text,
+        )
+
+
+class NavigationDemoGui:
+    def __init__(self, args):
+        self.args = args
+        self.engine: NavigationDemoEngine | None = None
+        self.running = False
+        self.main_texture_size: tuple[int, int] | None = None
+        self.snapshot_texture_size: tuple[int, int] | None = None
+        self.main_texture_tag = TEXTURE_MAIN
+        self.snapshot_texture_tag = TEXTURE_SNAPSHOT
+        self.last_state: DemoFrameState | None = None
+
+    def start(self) -> None:
+        if self.engine is None:
+            self.engine = NavigationDemoEngine(self.args)
+            dpg.set_value(TEXT_STATUS, f"Sorgente: {self.engine.source_name}")
+        self.running = True
+        dpg.configure_item(BUTTON_RUN, label="Pausa")
+
+    def pause(self) -> None:
+        self.running = False
+        dpg.configure_item(BUTTON_RUN, label="Avvia")
+
+    def toggle_run(self) -> None:
+        if self.running:
+            self.pause()
+        else:
+            self.start()
+
+    def reset(self) -> None:
+        self.pause()
+        if self.engine is not None:
+            self.engine.close()
+        self.engine = NavigationDemoEngine(self.args)
+        self.last_state = None
+        self.main_texture_size = None
+        self.snapshot_texture_size = None
+        dpg.set_value(TEXT_COMMANDS, "Nessun comando generato")
+        dpg.set_value(TEXT_TRACKS, "Nessuna track")
+        dpg.set_value(TEXT_METRICS, "Metriche non disponibili")
+        dpg.set_value(TEXT_STATUS, f"Reset completato. Sorgente: {self.engine.source_name}")
+
+    def step_once(self) -> None:
+        if self.engine is None:
+            self.engine = NavigationDemoEngine(self.args)
+        state = self.engine.step()
+        if state is None:
+            self.pause()
+            dpg.set_value(TEXT_STATUS, self.engine.stop_reason if self.engine else "Stop")
+            return
+        self.last_state = state
+        self._update_ui(state)
+
+    def tick(self) -> None:
+        if self.running:
+            self.step_once()
+
+    def close(self) -> None:
+        if self.engine is not None:
+            self.engine.close()
+
+    def _update_ui(self, state: DemoFrameState) -> None:
+        self._set_image_texture(
+            frame=state.display_frame,
+            image_tag=IMAGE_MAIN,
+            size_attr="main_texture_size",
+            tag_attr="main_texture_tag",
+            max_display_width=960,
+        )
+        if state.snapshot_frame is not None:
+            self._set_image_texture(
+                frame=nav.resize_preview(state.snapshot_frame, 520),
+                image_tag=IMAGE_SNAPSHOT,
+                size_attr="snapshot_texture_size",
+                tag_attr="snapshot_texture_tag",
+                max_display_width=520,
+            )
+
+        assert self.engine is not None
+        avg_yolo = self.engine.yolo_total_ms / max(self.engine.yolo_cycles, 1)
+        dpg.set_value(
+            TEXT_METRICS,
+            "\n".join([
+                state.fps_text,
+                f"YOLO ultimo: {state.yolo_ms:.1f} ms",
+                f"YOLO medio: {avg_yolo:.1f} ms",
+                self.engine.navigator.motion_text,
+                f"Snapshot salvati: {self.engine.navigator.snapshot_counter}",
+                f"Output: {Path(self.args.snapshot_output_dir).resolve()}",
+            ]),
+        )
+        dpg.set_value(
+            TEXT_COMMANDS,
+            "\n".join(self.engine.navigator.last_command_lines)
+            if self.engine.navigator.last_command_lines
+            else "Nessun comando generato",
+        )
+        dpg.set_value(TEXT_TRACKS, self._format_tracks(state.tracks))
+        if state.snapshots:
+            dpg.set_value(
+                TEXT_STATUS,
+                f"Snapshot {state.snapshots[-1].snapshot_id:04d} acquisito "
+                f"su {state.snapshots[-1].simulated_position}",
+            )
+
+    def _set_image_texture(
+        self,
+        frame: np.ndarray,
+        image_tag: str,
+        size_attr: str,
+        tag_attr: str,
+        max_display_width: int,
+    ) -> None:
+        h, w = frame.shape[:2]
+        rgba = bgr_to_rgba_float(frame)
+        display_scale = min(1.0, max_display_width / float(w))
+        display_w = max(1, int(w * display_scale))
+        display_h = max(1, int(h * display_scale))
+        old_size = getattr(self, size_attr)
+        if old_size != (w, h):
+            old_texture_tag = getattr(self, tag_attr)
+            new_texture_tag = dpg.generate_uuid()
+            dpg.add_dynamic_texture(
+                w,
+                h,
+                rgba,
+                tag=new_texture_tag,
+                parent=TEXTURE_REGISTRY,
+            )
+            dpg.configure_item(
+                image_tag,
+                texture_tag=new_texture_tag,
+                width=display_w,
+                height=display_h,
+            )
+            if dpg.does_item_exist(old_texture_tag):
+                dpg.delete_item(old_texture_tag)
+            setattr(self, tag_attr, new_texture_tag)
+            setattr(self, size_attr, (w, h))
+            return
+        dpg.configure_item(image_tag, width=display_w, height=display_h)
+        dpg.set_value(getattr(self, tag_attr), rgba)
+
+    @staticmethod
+    def _format_tracks(tracks: list[nav.Track]) -> str:
+        if not tracks:
+            return "Nessuna track"
+        lines = []
+        for track in tracks[:12]:
+            cx, cy = nav.bbox_center(track.bbox)
+            lines.append(
+                f"#{track.id:03d} {track.state:<11} "
+                f"conf={track.confidence:.2f} hits={track.hits:<3} "
+                f"missed={track.missed:<2} center=({cx:.0f},{cy:.0f}) "
+                f"trend={track.area_trend():+.2f} {track.last_candidate_reason}"
+            )
+        return "\n".join(lines)
+
+
+def bgr_to_rgba_float(frame: np.ndarray) -> np.ndarray:
+    rgba = cv2.cvtColor(frame, cv2.COLOR_BGR2RGBA)
+    return np.asarray(rgba, dtype=np.float32).ravel() / 255.0
+
+
+def build_ui(app: NavigationDemoGui) -> None:
+    dpg.create_context()
+
+    with dpg.font_registry():
+        default_font = dpg.add_font("C:/Windows/Fonts/segoeui.ttf", 17)
+        mono_font = dpg.add_font("C:/Windows/Fonts/consola.ttf", 15)
+
+    with dpg.texture_registry(show=False, tag=TEXTURE_REGISTRY):
+        blank = np.zeros((32, 32, 4), dtype=np.float32).ravel()
+        dpg.add_dynamic_texture(32, 32, blank, tag=TEXTURE_MAIN)
+        dpg.add_dynamic_texture(32, 32, blank, tag=TEXTURE_SNAPSHOT)
+
+    with dpg.window(tag="main_window", label="FlyWMS Navigation Demo"):
+        with dpg.group(horizontal=True):
+            dpg.add_button(label="Avvia", tag=BUTTON_RUN, callback=lambda: app.toggle_run())
+            dpg.add_button(label="Step", callback=lambda: app.step_once())
+            dpg.add_button(label="Reset", callback=lambda: app.reset())
+            dpg.add_text("Pronto", tag=TEXT_STATUS)
+
+        dpg.add_separator()
+
+        with dpg.group(horizontal=True):
+            with dpg.child_window(width=980, height=-1, border=False):
+                dpg.add_image(TEXTURE_MAIN, tag=IMAGE_MAIN)
+
+            with dpg.child_window(width=560, height=-1, border=False):
+                dpg.add_text("Metriche")
+                dpg.add_text("Metriche non disponibili", tag=TEXT_METRICS)
+                dpg.bind_item_font(TEXT_METRICS, mono_font)
+                dpg.add_separator()
+                dpg.add_text("Comandi")
+                dpg.add_text("Nessun comando generato", tag=TEXT_COMMANDS)
+                dpg.bind_item_font(TEXT_COMMANDS, mono_font)
+                dpg.add_separator()
+                dpg.add_text("Payload OCR")
+                dpg.add_image(TEXTURE_SNAPSHOT, tag=IMAGE_SNAPSHOT)
+                dpg.add_separator()
+                dpg.add_text("Track")
+                dpg.add_text("Nessuna track", tag=TEXT_TRACKS)
+                dpg.bind_item_font(TEXT_TRACKS, mono_font)
+
+    dpg.bind_font(default_font)
+    dpg.create_viewport(title="FlyWMS Navigation Demo", width=1600, height=980)
+    dpg.setup_dearpygui()
+    dpg.show_viewport()
+    dpg.set_primary_window("main_window", True)
+
+
+def main() -> int:
+    if dpg is None:
+        print(
+            "DearPyGUI non e' installato. Installa il pacchetto con:\n"
+            "  python -m pip install dearpygui\n"
+            f"Dettaglio import: {DEARPYGUI_IMPORT_ERROR}",
+            flush=True,
+        )
+        return 1
+
+    args = nav.parse_args()
+    args.no_display = True
+    app = NavigationDemoGui(args)
+    build_ui(app)
+    try:
+        while dpg.is_dearpygui_running():
+            app.tick()
+            dpg.render_dearpygui_frame()
+    finally:
+        app.close()
+        dpg.destroy_context()
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
diff --git a/handoff.md b/handoff.md
new file mode 100644
index 0000000..27ff2e7
--- /dev/null
+++ b/handoff.md
@@ -0,0 +1,116 @@
+# FlyWMS handoff - GUI navigazione
+
+Data: 2026-05-15
+
+## Stato repository
+
+- Branch: `master`
+- Remote: `origin` su Gitea
+- Ultimo lavoro: aggiunta shell DearPyGUI separata per la simulazione dimostrativa di `flywms_navigation.py`.
+
+## Contesto recuperato
+
+L'obiettivo discusso era trasformare `flywms_navigation.py` in una simulazione piu' ordinata e dimostrativa, senza contaminare la logica esistente. Il programma originale contiene gia':
+
+- detector YOLO/Ultralytics;
+- tracking geometrico leggero dei `gaylord`;
+- decisione di snapshot quando una track e' centrata;
+- salvataggio frame debug, payload OCR e `snapshots.jsonl`;
+- simulazione comandi `STOP`, `SCATTA_FOTO`, `ASSOCIA_POSIZIONE`, `INVIA_ROI_REMOTA`, `ATTENDI_ACK`, `RIPARTI_*`;
+- UI debug OpenCV integrata nel `main()`.
+
+## Decisione architetturale
+
+La GUI DearPyGUI e' stata tenuta in un file separato, `flywms_navigation_gui.py`.
+
+Motivo: non toccare il ciclo originale e non rischiare regressioni nella logica di navigazione. La GUI importa `flywms_navigation.py` e riusa:
+
+- `UltralyticsDetector`;
+- `LightweightTracker`;
+- `NavigationController`;
+- `draw_navigation_debug`;
+- configurazione da `flywms_navigation.ini`.
+
+In pratica `flywms_navigation.py` resta il simulatore CLI/OpenCV funzionante; `flywms_navigation_gui.py` e' una shell dimostrativa sopra la stessa logica.
+
+## Implementato
+
+File nuovo:
+
+- `flywms_navigation_gui.py`
+
+Funzioni principali:
+
+- `NavigationDemoEngine`: esegue un singolo step del ciclo video/YOLO/tracking/navigazione e restituisce uno stato frame-by-frame.
+- `NavigationDemoGui`: gestisce finestra DearPyGUI, comandi Avvia/Pausa, Step, Reset, preview, metriche, comandi, payload OCR e track.
+- conversione frame OpenCV BGR in texture RGBA DearPyGUI.
+- gestione texture corretta con tag univoci quando cambia la dimensione immagine.
+- resize visuale della preview principale a 960x540.
+
+DearPyGUI e' stato installato nell'ambiente Python corrente con:
+
+```powershell
+python -m pip install dearpygui
+```
+
+## Verifiche eseguite
+
+Compilazione:
+
+```powershell
+python -m py_compile flywms_navigation_gui.py flywms_navigation.py
+```
+
+Test motore GUI senza finestra persistente:
+
+```text
+state1 1 (720, 1280, 3)
+state2 2 (720, 1280, 3)
+```
+
+Test aggiornamento UI/texture:
+
+```text
+ui step ok
+960 540
+```
+
+Avvio manuale usato:
+
+```powershell
+python flywms_navigation_gui.py
+```
+
+## Osservazioni performance
+
+Con CPU, la GUI scende circa da 6 fps a 3.8 fps. Il dato e' credibile perche' la shell GUI aggiunge:
+
+- disegno overlay OpenCV;
+- conversione BGR -> RGBA;
+- normalizzazione/copia texture `float32`;
+- aggiornamento pannelli DearPyGUI nello stesso thread di YOLO.
+
+Il carico CPU alto su tutti gli 8 core e' credibile: PyTorch/OpenCV/NumPy usano thread nativi anche se il loop Python e' singolo.
+
+Quando sara' disponibile una GPU corretta, conviene rimisurare separando:
+
+- fps YOLO/tracking;
+- fps GUI;
+- tempo medio YOLO;
+- tempo medio conversione texture.
+
+## Prossimi passi consigliati
+
+1. Migliorare la GUI come demo:
+   - timeline eventi `STOP`, `SCATTA_FOTO`, `ACK/NACK`, `RIPARTI`;
+   - pannello soglie principali;
+   - evidenza piu' leggibile dello stato `entering/candidate/centered/snapshotted`;
+   - contatori snapshot e posizioni simulate.
+
+2. Ottimizzare solo se serve:
+   - aggiornare texture a frequenza limitata;
+   - convertire per GUI un frame gia' ridotto;
+   - separare inferenza e GUI con stato latest-frame;
+   - aggiornare testo solo quando cambia.
+
+3. Non modificare la logica di `flywms_navigation.py` finche' la GUI non ha stabilizzato i requisiti dimostrativi.