diff --git a/airc b/airc
index 2339e6d..189dad2 100755
--- a/airc
+++ b/airc
@@ -291,6 +291,22 @@ fi
 unset _gh_resolved
 
 AIRC_WRITE_DIR="$(detect_scope)"
+
+# Write a sentinel marker before any intentional `exec env ... "$0" ...`
+# call, so the Windows daemon launcher .bat can distinguish "intentional
+# re-exec into different mode" from "actual crash" (#203). On Linux/Mac
+# `exec` is a true execve — the parent bash's PID becomes the new
+# program, so the launcher script never observes an exit and the marker
+# is harmless. On Windows MSYS-bash, exec is emulated as spawn-and-exit:
+# the original bash exits + a new airc bash takes over. The launcher
+# .bat sees the original bash exit, would normally treat it as a crash,
+# and respawn — racing the new airc that just took over (Joel/continuum-
+# b69f's #203 crashloop). Marker contents: "PID:UNIX_TIMESTAMP". Caller
+# is responsible for invoking this immediately before exec.
+_write_reexec_marker() {
+  local marker="$AIRC_WRITE_DIR/airc.reexec-marker"
+  printf '%d:%d\n' "$$" "$(date +%s)" > "$marker" 2>/dev/null || true
+}
 CONFIG="$AIRC_WRITE_DIR/config.json"
 IDENTITY_DIR="$AIRC_WRITE_DIR/identity"
 PEERS_DIR="$AIRC_WRITE_DIR/peers"
@@ -2182,11 +2198,13 @@ cmd_connect() {
       if [ -n "$_new_picked" ]; then
         echo "  ✓ Another tab beat us to it — joining their fresh gist ($_new_picked)"
         echo ""
+        _write_reexec_marker
         exec env ${_preserved_name:+AIRC_NAME="$_preserved_name"} "$0" connect "$_new_picked"
       fi
 
       echo "  Re-execing into host mode for #${resolved_room_name}..."
       echo ""
+      _write_reexec_marker
       exec env AIRC_NO_DISCOVERY=1 ${_preserved_name:+AIRC_NAME="$_preserved_name"} "$0" connect --room "$resolved_room_name"
     fi
 
@@ -2386,6 +2404,7 @@ except Exception:
           echo "  ✓ Another tab beat us to it — joining their fresh gist ($_new_picked)"
           echo ""
           # Re-exec as joiner pointing at the winner's gist.
+          _write_reexec_marker
           exec env ${_preserved_name:+AIRC_NAME="$_preserved_name"} "$0" connect "$_new_picked"
         fi
 
@@ -2394,6 +2413,7 @@ except Exception:
         # exec replaces the current bash process. AIRC_NO_DISCOVERY=1
         # prevents the new instance from re-finding the just-deleted gist
         # (gh's gist-list cache might still show it for a few seconds).
+        _write_reexec_marker
         exec env AIRC_NO_DISCOVERY=1 ${_preserved_name:+AIRC_NAME="$_preserved_name"} "$0" connect --room "$resolved_room_name"
       fi
       # Either not a room flow, or no gh, or no resolved_room_name → original die.
@@ -2834,6 +2854,7 @@ JSON
                       "$AIRC_WRITE_DIR/room_gist_id" \
                       "$AIRC_WRITE_DIR/room_name"
                 local _preserved_name; _preserved_name=$(get_config_val name "")
+                _write_reexec_marker
                 exec env ${_preserved_name:+AIRC_NAME="$_preserved_name"} "$0" connect "$_winner_id"
               fi
             fi
@@ -5046,15 +5067,46 @@ _daemon_install_schtasks() {
     cwd_win=$(printf '%s' "$(pwd -P)" | sed 's|^/\([a-z]\)/|\U\1:\\\\|; s|/|\\\\|g')
     airc_bin_unix="$airc_bin"
   fi
+  # Marker path the daemon-launcher polls between iterations to
+  # distinguish "intentional re-exec into different mode" from "actual
+  # crash" (#203). airc itself writes this file via _write_reexec_marker
+  # right before any `exec env ... "$0" connect ...` call. On Windows
+  # MSYS-bash, exec is emulated as spawn-and-exit (not a true execve),
+  # so the launcher .bat sees the original bash exit while the new
+  # airc takes over — the marker tells the .bat to step aside instead
+  # of racing-respawn the new airc with another instance.
+  local marker_win
+  if command -v cygpath >/dev/null 2>&1; then
+    marker_win=$(cygpath -w "$scope/airc.reexec-marker")
+  else
+    marker_win=$(printf '%s' "$scope/airc.reexec-marker" | sed 's|^/\([a-z]\)/|\U\1:\\\\|; s|/|\\\\|g')
+  fi
   local launcher_bash="$scope/airc-daemon.bat"
   cat > "$launcher_bash" <<EOF
 @echo off
 REM AIRC daemon launcher — generated by 'airc daemon install' on Windows.
 REM Runs airc connect under bash, restarting on exit. Logs to daemon.log.
+REM On intentional re-exec (host-takeover or rejoin-as-joiner), airc
+REM writes airc.reexec-marker — we step aside rather than respawn,
+REM since the new airc bash from the exec is now the daemon.
 cd /d "$cwd_win"
 set AIRC_BACKGROUND_OK=1
 :loop
 "$bash_exe" -c "exec '$airc_bin_unix' connect"
+REM Did airc just intentionally re-exec? If marker exists and is recent,
+REM the new airc process from the exec is now the running daemon —
+REM exit the launcher loop instead of racing-respawn it.
+REM forfiles /m airc.reexec-marker /d 0 /c "cmd /c exit 0" succeeds when
+REM the file's mtime is today (fine-grained age check below via type +
+REM date math is too brittle for .bat; "today" is our 60s proxy).
+if exist "$marker_win" (
+  forfiles /p "$scope_win" /m airc.reexec-marker /d 0 /c "cmd /c exit 0" >nul 2>&1
+  if not errorlevel 1 (
+    echo [%date% %time%] airc re-exec'd into different mode ^(host-takeover or rejoin^); new process is now daemon, launcher exiting. >> daemon.err
+    del "$marker_win" >nul 2>&1
+    exit /b 0
+  )
+)
 echo [%date% %time%] airc connect exited. Restarting in 5s. >> daemon.err
 timeout /t 5 /nobreak >nul
 goto loop