feat: TTS backend abstraction + 4 'imba' features + Lua packs

Big push driven by user feedback ("делай имбу") and web research on what
voice assistants need to be the ideal:

TTS backend abstraction (P0.1)
  - new module crates/jarvis-core/src/tts/{mod,sapi,piper,silero}.rs
  - TtsBackend trait with SapiBackend (current PowerShell), PiperBackend
    (rhasspy/piper, neural quality), SileroBackend (python subprocess)
  - JARVIS_TTS env var picks (sapi|piper|silero). Auto-detect Piper if
    binary + voice present in tools/piper/. Falls back to SAPI on missing.
  - SpeakOpts {lang, detached, raw} replaces ad-hoc args. text_utils
    sanitiser applied unless raw=true.
  - llm_fallback + lua/api/tts both routed through tts::backend().
  - tools/piper/install.ps1 downloads piper.exe + ru_RU-irina-medium.onnx
    from rhasspy releases + huggingface. Smoke-test included.
  - tools/silero/silero_tts.py helper (PyTorch); rust spawns it as subprocess.

IMBA-1 Agentic LLM router
  - crates/jarvis-app/src/llm_router.rs
  - When fuzzy/intent matcher fails, LLM picks the closest command from the
    full registry. Returns JSON {command_id, confidence, reason}.
  - Threshold-gated re-dispatch via substitute phrase. JARVIS_LLM_ROUTER=1
    enables; JARVIS_LLM_ROUTER_THRESHOLD overrides 0.55 default.
  - Inserted in app.rs::execute_command between "no match" and existing
    llm_fallback chat fallback.

IMBA-2 Long-term memory
  - crates/jarvis-core/src/long_term_memory.rs — JSON store at
    APP_CONFIG_DIR/long_term_memory.json. Atomic write-through.
  - remember/recall/search/forget/all/build_context API.
  - Lua bindings: jarvis.memory.* (5 functions).
  - llm_fallback auto-injects relevant facts (substring search of prompt)
    into system message before LLM call.
  - Pack resources/commands/memory_pack/ with 4 commands: remember, recall,
    forget, list.

IMBA-3 Profile switching (work/game/sleep/driving/default)
  - crates/jarvis-core/src/profiles.rs — JSON profiles at APP_CONFIG_DIR/profiles/
    Auto-seeds 5 defaults on first run with personality + allow/deny lists +
    greetings + emoji icons.
  - active_profile.txt persists choice across restart.
  - Lua bindings: jarvis.profile.{active,set,list,allows,active_name}.
  - llm_fallback prepends profile personality to system prompt.
  - Pack resources/commands/profile_switch/ with 6 voice triggers.

IMBA-4 Multimodal screenshot + vision LLM
  - crates/jarvis-core/src/lua/api/vision.rs — gated on HTTP sandbox.
  - jarvis.vision.screenshot() captures via PowerShell System.Drawing.
  - jarvis.vision.describe(prompt?) sends base64 PNG to Groq vision model
    (default llama-3.2-11b-vision-preview, override via GROQ_VISION_MODEL).
  - Pack resources/commands/vision/ with 2 commands: describe + read_error.

P0.2 Continuous conversation grace window
  - config::CONVERSATION_GRACE_MS = 30_000.
  - app.rs: after command result, if grace_ms > 0 keep listening WITHOUT
    re-wake for the grace duration. Existing CMS_WAIT_DELAY back-dated so
    the existing timeout fires at start + grace_ms.

Tests: 24/24 jarvis-core unit tests pass (including 5 text_utils).
Build: cargo build --release -p jarvis-app and -p jarvis-gui both succeed
on Windows MSVC (VS 2026 Enterprise vcvars64).

Notes for setup:
  - Piper voice install: pwsh tools/piper/install.ps1 (downloads ~90 MB).
  - GROQ_TOKEN needed for IMBA-1 (router) and IMBA-4 (vision).
  - All features are opt-in via env vars or auto-detect; existing SAPI +
    fuzzy match path remains the default.

tools/piper/README.md

@@ -0,0 +1,64 @@
+# Piper TTS (neural voice for J.A.R.V.I.S.)
+
+Drop-in replacement for the robotic SAPI voice. ~30 MB binary + ~60 MB Russian voice (Irina). Latency 200-500ms on i5.
+
+## Install
+
+```powershell
+cd C:\Jarvis\rust\tools\piper
+pwsh -File install.ps1
+```
+
+By default installs `ru_RU-irina-medium` (Russian female). To pick a different voice:
+
+```powershell
+pwsh -File install.ps1 -Voice ru_RU-denis-medium
+pwsh -File install.ps1 -Voice ru_RU-dmitri-medium
+pwsh -File install.ps1 -Voice ru_RU-ruslan-medium
+pwsh -File install.ps1 -Voice en_US-amy-medium      # English female
+```
+
+Force re-download:
+
+```powershell
+pwsh -File install.ps1 -Force
+```
+
+## Use
+
+If `piper.exe` + a voice file are present in `tools/piper/`, J.A.R.V.I.S. auto-detects Piper on startup. No env vars required.
+
+To force a specific backend:
+
+```bat
+set JARVIS_TTS=piper      :: use Piper
+set JARVIS_TTS=sapi       :: force SAPI fallback
+set JARVIS_TTS=silero     :: use Silero (needs tools/silero/silero_tts.py)
+```
+
+Override paths:
+
+```bat
+set JARVIS_TTS_PIPER_BIN=D:\custom\piper.exe
+set JARVIS_TTS_PIPER_VOICE=D:\voices\my-voice.onnx
+```
+
+## Voices catalogue
+
+Full list: https://huggingface.co/rhasspy/piper-voices
+
+Recommended for Russian:
+- `ru_RU-irina-medium` — natural female, default
+- `ru_RU-dmitri-medium` — male, clear diction
+- `ru_RU-ruslan-medium` — male, slightly deeper
+- `ru_RU-denis-medium` — male, "newsreader" tone
+
+## Troubleshooting
+
+**"piper.exe not found"** — install script failed; check `install.ps1` output. The fallback is SAPI — assistant still works, just with the robotic voice.
+
+**"no .onnx voice in voices/"** — voice download failed (HuggingFace blocked?). Manually download from https://huggingface.co/rhasspy/piper-voices/tree/main/ru/ru_RU/irina/medium and drop both `.onnx` + `.onnx.json` into `voices/`.
+
+**Voice sounds wrong** — try a different model. Each Russian voice has a different timbre.
+
+**No audio output** — Piper writes to a temp wav, then plays via `System.Media.SoundPlayer` synchronously. Check that the default playback device isn't muted.

tools/piper/install.ps1

@@ -0,0 +1,108 @@
+#requires -Version 5.1
+<#
+.SYNOPSIS
+  Download Piper TTS engine + Russian voice into tools/piper/.
+
+.DESCRIPTION
+  Fetches the latest piper_windows_amd64.zip from rhasspy/piper-binary and
+  the Irina (ru_RU) voice .onnx + .json from huggingface rhasspy/piper-voices.
+  Idempotent — skips if files already exist. Pass -Force to re-download.
+
+.PARAMETER Voice
+  Voice name to install. Default "ru_RU-irina-medium". Other options:
+  ru_RU-denis-medium, ru_RU-dmitri-medium, ru_RU-ruslan-medium, en_US-amy-medium, ...
+
+.PARAMETER Force
+  Re-download even if files exist.
+
+.EXAMPLE
+  pwsh -File install.ps1
+  pwsh -File install.ps1 -Voice ru_RU-denis-medium
+  pwsh -File install.ps1 -Force
+#>
+
+param(
+    [string]$Voice = "ru_RU-irina-medium",
+    [switch]$Force
+)
+
+$ErrorActionPreference = "Stop"
+$InformationPreference = "Continue"
+
+$Root = Split-Path -Parent $MyInvocation.MyCommand.Path
+$BinPath = Join-Path $Root "piper.exe"
+$VoiceDir = Join-Path $Root "voices"
+$VoiceFile = Join-Path $VoiceDir "$Voice.onnx"
+$VoiceJson = Join-Path $VoiceDir "$Voice.onnx.json"
+
+# ─── Piper binary ──────────────────────────────────────────────────────────
+if ((Test-Path $BinPath) -and -not $Force) {
+    Write-Information "[skip] piper.exe already present at $BinPath"
+} else {
+    Write-Information "[fetch] piper_windows_amd64.zip..."
+    $zip = Join-Path $env:TEMP "piper_windows_amd64.zip"
+    $url = "https://github.com/rhasspy/piper/releases/latest/download/piper_windows_amd64.zip"
+    Invoke-WebRequest -Uri $url -OutFile $zip -UseBasicParsing
+    Write-Information "[unzip] -> $Root"
+    Expand-Archive -Path $zip -DestinationPath $Root -Force
+    # piper distributes inside a "piper/" subfolder — flatten it
+    $inner = Join-Path $Root "piper"
+    if (Test-Path $inner) {
+        Get-ChildItem -Path $inner -Force | Move-Item -Destination $Root -Force
+        Remove-Item -Path $inner -Recurse -Force
+    }
+    Remove-Item $zip -Force
+    if (-not (Test-Path $BinPath)) {
+        throw "piper.exe not found after extraction"
+    }
+    Write-Information "[ok] piper installed at $BinPath"
+}
+
+# ─── Voice files ───────────────────────────────────────────────────────────
+if (-not (Test-Path $VoiceDir)) {
+    New-Item -ItemType Directory -Path $VoiceDir | Out-Null
+}
+
+# Voice naming: ru_RU-irina-medium -> lang=ru_RU, name=irina, quality=medium
+$parts = $Voice -split "-"
+if ($parts.Length -lt 3) {
+    throw "Bad voice name '$Voice' — expected <lang>-<name>-<quality>"
+}
+$lang = $parts[0]
+$name = $parts[1]
+$quality = $parts[2]
+$langCode = ($lang -split "_")[0]
+
+$Base = "https://huggingface.co/rhasspy/piper-voices/resolve/main/$langCode/$lang/$name/$quality"
+
+foreach ($pair in @(
+    @{ url = "$Base/$Voice.onnx"; out = $VoiceFile },
+    @{ url = "$Base/$Voice.onnx.json"; out = $VoiceJson }
+)) {
+    if ((Test-Path $pair.out) -and -not $Force) {
+        Write-Information "[skip] $(Split-Path -Leaf $pair.out) already present"
+        continue
+    }
+    Write-Information "[fetch] $($pair.url)"
+    Invoke-WebRequest -Uri $pair.url -OutFile $pair.out -UseBasicParsing
+    Write-Information "[ok] $(Split-Path -Leaf $pair.out) -> $($pair.out)"
+}
+
+# ─── Smoke test ────────────────────────────────────────────────────────────
+$testWav = Join-Path $env:TEMP "jarvis-piper-test.wav"
+try {
+    Write-Information "[test] synthesizing 'Привет, Дмитрий'..."
+    "Привет, Дмитрий" | & $BinPath --model $VoiceFile --output_file $testWav 2>$null
+    if ((Test-Path $testWav) -and ((Get-Item $testWav).Length -gt 1000)) {
+        Write-Information "[ok] Piper works! Test wav: $testWav"
+        Write-Information ""
+        Write-Information "Next: set environment variable JARVIS_TTS=piper"
+        Write-Information "      (or just leave it unset — auto-detect will pick Piper)"
+    } else {
+        Write-Warning "Test wav empty or missing — synthesis may have failed"
+    }
+} catch {
+    Write-Warning "Smoke test failed: $_"
+} finally {
+    if (Test-Path $testWav) { Remove-Item $testWav -Force }
+}

tools/silero/silero_tts.py

@@ -0,0 +1,85 @@
+"""Silero TTS helper for J.A.R.V.I.S.
+
+Reads text from stdin, synthesises with snakers4/silero-models, writes the result
+to a temp wav, and prints the path on stdout. The Rust side then plays + deletes it.
+
+Install:
+    pip install torch soundfile numpy
+
+Usage (called by jarvis-core/src/tts/silero.rs):
+    echo "привет" | python silero_tts.py --voice xenia
+
+Voices for ru_v3:
+    xenia (female, default), baya (female), aidar (male), eugene (male), kseniya (female).
+"""
+
+import argparse
+import os
+import sys
+import tempfile
+
+try:
+    import torch  # type: ignore
+except ImportError:
+    print("silero_tts: torch not installed. Run `pip install torch`.", file=sys.stderr)
+    sys.exit(2)
+
+import warnings
+
+warnings.filterwarnings("ignore")
+
+_MODEL = None
+
+
+def load_model(device: str = "cpu"):
+    global _MODEL
+    if _MODEL is None:
+        torch.set_num_threads(4)
+        _MODEL, _ = torch.hub.load(
+            repo_or_dir="snakers4/silero-models",
+            model="silero_tts",
+            language="ru",
+            speaker="v3_1_ru",
+            trust_repo=True,
+        )
+        _MODEL.to(torch.device(device))
+    return _MODEL
+
+
+def synth(text: str, voice: str, sample_rate: int = 48000) -> str:
+    """Synth text → return path to temp wav."""
+    model = load_model()
+    audio = model.apply_tts(text=text, speaker=voice, sample_rate=sample_rate)
+
+    import numpy as np
+    import soundfile as sf  # type: ignore
+
+    fd, path = tempfile.mkstemp(prefix="jarvis-silero-", suffix=".wav")
+    os.close(fd)
+    sf.write(path, audio.numpy().astype(np.float32), sample_rate)
+    return path
+
+
+def main() -> int:
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--voice", default="xenia", help="speaker name (xenia, baya, aidar, eugene, kseniya)")
+    ap.add_argument("--sample-rate", type=int, default=48000)
+    args = ap.parse_args()
+
+    text = sys.stdin.read().strip()
+    if not text:
+        print("silero_tts: empty input", file=sys.stderr)
+        return 1
+
+    try:
+        path = synth(text, args.voice, args.sample_rate)
+    except Exception as exc:
+        print(f"silero_tts: synth failed: {exc}", file=sys.stderr)
+        return 3
+
+    print(path)
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())