diff --git a/CLAUDE.md b/CLAUDE.md
index 61c2fbe..b23f4db 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -33,18 +33,23 @@ Frontend, desktop, and tooling commands live in their nested CLAUDE.md.
 ```
 YAML scenario → compile_runtime3.py → Runtime 3 IR → ESP32 / Web player
                                                     ↓
-                                              Voice bridge → Piper TTS (Tower:8001)
+                                              Voice bridge :8200 (MacStudio)
+                                                ↳ /tts → F5-TTS-MLX (cache → F5 → safe wav)
+                                                ↳ /voice/intent → LiteLLM npc-fast (MLX 7B Q4)
+                                                ↳ /voice/transcribe → whisper.cpp :8300
                                                     ↓
-                                              Hints engine → /hints/ask
+                                              Hints engine → /hints/ask (LLM rewrite via hints-deep MLX 32B Q4)
 ```
 
 Key surfaces:
 - **Scenario IR**: `game/scenarios/zacus_v2.yaml` → `tools/scenario/compile_runtime3.py` → portable Runtime 3 IR. Contract: `specs/ZACUS_RUNTIME_3_SPEC.md`.
 - **Authoring**: `frontend-v3/` (pnpm monorepo: `apps/atelier/` Scratch-like studio + `apps/dashboard/` game-master live view).
 - **Firmware**:
-  - `ESP32_ZACUS/` submodule — Freenove ESP32-S3 master (NPC engine, voice pipeline, vision/QR, media manager).
+  - `ESP32_ZACUS/` submodule — Freenove ESP32-S3 master (NPC engine, voice pipeline, vision/QR, media manager). Branche `feat/idf-migration` en cours (port Arduino → ESP-IDF 5.4).
   - `PLIP_FIRMWARE/` — retro telephone annex (ES8388 dev kit bringup, Si3210 PCB target). REST endpoint consumed by the master ESP32.
-- **Voice / NPC**: Piper TTS on Tower:8001 (zacus voice = tom-medium). NPC phrases in `game/scenarios/npc_phrases.yaml`. MP3 pool generator: `tools/tts/generate_npc_pool.py`.
+- **Voice / NPC**: F5-TTS-MLX in-process sur MacStudio (`studio:8200`), voix Zacus clonée via `~/zacus_reference.wav`. Pré-rendu via `tools/tts/generate_npc_pool.py --backend f5`. Service down WAV hardcoded en graceful degradation. NPC phrases : `game/scenarios/npc_phrases.yaml`.
+- **Hints engine**: `tools/hints/server.py` FastAPI (LLM rewrite via LiteLLM `hints-deep`, anti-cheat, adaptive level, safety filter `game/scenarios/hints_safety.yaml`).
+- **Inference stack**: voir `tools/macstudio/README.md` — MLX-LM (npc + hints), whisper.cpp, F5-TTS, LiteLLM proxy `:4000`. Configs versionnées dans `tools/macstudio/`.
 - **MCP hardware**: `tools/dev/mcp_hardware_server.py` (stdio, 6 tools).
 - **Desktop hub**: `desktop/` (Electron, macOS, bundles V3 frontends, talks USB serial).
 
@@ -76,8 +81,9 @@ Key surfaces:
 
 ## Infrastructure
 
-- **Tower** (`clems@192.168.0.120`): Piper TTS FR on `:8001` (voices: tom-medium / siwis / upmc).
-- **KXKM-AI** (`kxkm@kxkm-ai` via Tailscale): RTX 4090, GPU inference.
+- **MacStudio** (`ssh studio`, M3 Ultra 512 GB) : **stack inférence Zacus complète** — MLX-LM (`npc-fast` 7B Q4 :8501, `hints-deep` 32B Q4 :8500), whisper.cpp `large-v3-turbo-q5_0` :8300, F5-TTS-MLX in-process via voice-bridge :8200 (cache disque + watchdog `*/2` + `service_down.wav` graceful), LiteLLM proxy :4000. Voir `tools/macstudio/README.md`.
+- **Tower** (`clems@192.168.0.120`) : Suite Numérique, n8n, LiteLLM Tower (legacy). **Plus dans la chaîne live Zacus** ; container Docker `zacus-tts-piper` survit pour batch historique uniquement (voix EN-only, FR à déployer si besoin).
+- **KXKM-AI** (`kxkm@kxkm-ai` via Tailscale) : RTX 4090, GPU inference (fine-tuning, plus utilisé en runtime).
 - SSH is key-based only — never `sshpass`.
 
 ## Language
diff --git a/tools/macstudio/README.md b/tools/macstudio/README.md
index 72bc47b..eb294c5 100644
--- a/tools/macstudio/README.md
+++ b/tools/macstudio/README.md
@@ -1,30 +1,81 @@
 # MacStudio inference stack — config sources of truth
 
-Deployment configs for the MacStudio M3 Ultra (`ssh studio`) hosting Ollama,
+Deployment configs for the MacStudio M3 Ultra (`ssh studio`,
+`clems@MacStudio-de-MonsieurB.local`, 512 GB) hosting MLX-LM, Ollama,
 whisper.cpp, F5-TTS, and the LiteLLM proxy. Spec:
 `docs/superpowers/specs/2026-05-03-tts-stt-llm-macstudio-design.md`.
 
+## Stack live (post-P1 part1-9)
+
+Inference and voice stack runs entirely on studio. Tower no longer participates
+in the live Zacus runtime path.
+
+| Service | Port | Backend | LiteLLM alias(es) | Monitoring |
+|---------|------|---------|-------------------|------------|
+| MLX-LM 7B Q4 | `:8501` | `mlx_lm.server` | `npc-fast` (~12 tok/s warm) | `curl -s http://studio:8501/v1/models` |
+| MLX-LM 32B Q4 | `:8500` | `mlx_lm.server` | `hints-deep` (~4.4 tok/s warm) | `curl -s http://studio:8500/v1/models` |
+| Ollama | `:11434` | `ollama serve` | `npc-ollama-7b`, `hints-ollama-72b` (legacy A/B) | `curl -s http://studio:11434/api/tags` |
+| whisper.cpp | `:8300` | `whisper-server` (`large-v3-turbo-q5_0`) | `whisper-fr`, `stt-fr` | `pgrep -fl whisper-server` |
+| F5-TTS-MLX | `:8200` (in-process) | voice-bridge `/tts` + `/voice/{transcribe,intent,ws}` | `tts-zacus-primary` | `curl -s http://studio:8200/healthz` |
+| LiteLLM proxy | `:4000` | `litellm` | (master_key `sk-zacus-local-dev-do-not-share`) | `curl -s http://studio:4000/health/readiness` |
+
 ## Files
 
 | Local path | Remote path on studio | Purpose |
 |------------|-----------------------|---------|
-| `litellm-config.yaml` | `~/litellm-config/config.yaml` | LiteLLM proxy (`:4000`) — 6 model aliases (`npc-fast`, `hints-deep`, `whisper-fr`, `stt-fr`, `tts-zacus-primary`, `tts-zacus-fallback`) |
-| `voice-bridge/main.py` | `~/voice-bridge/main.py` | FastAPI voice-bridge scaffold (`:8200`) — routes `/voice/transcribe` (whisper :8300), `/voice/intent` (LiteLLM `npc-fast`), `/tts` (F5 primary, Piper Tower:8001 fallback) |
-| `voice-bridge/requirements.txt` | `~/voice-bridge/requirements.txt` | FastAPI deps (fastapi, httpx, uvicorn) |
+| `litellm-config.yaml` | `~/litellm-config/config.yaml` | LiteLLM proxy (`:4000`) — 6+ model aliases (`npc-fast`, `hints-deep`, `whisper-fr`, `stt-fr`, `tts-zacus-primary`, `npc-ollama-7b`, `hints-ollama-72b`) |
+| `voice-bridge/main.py` | `~/voice-bridge/main.py` | FastAPI voice-bridge (`:8200`) — routes `/voice/transcribe` (whisper :8300), `/voice/intent` (LiteLLM `npc-fast`), `/tts` (F5-TTS-MLX in-process), `/voice/ws` |
+| `voice-bridge/requirements.txt` | `~/voice-bridge/requirements.txt` | FastAPI deps (fastapi, httpx, uvicorn, f5-tts-mlx) |
+| `voice-bridge/watchdog.sh` | `~/voice-bridge/watchdog.sh` | Watchdog: relance voice-bridge si `pgrep` absent |
 
-## Persistence (headless macOS — no launchd)
+## Aliases LiteLLM
+
+Model routing exposed via the proxy on `:4000`. Clients hit aliases (never raw
+backends) so we can swap MLX ↔ Ollama without code changes.
+
+| Alias | Backend | Use case |
+|-------|---------|----------|
+| `npc-fast` | MLX-LM 7B Q4 (`:8501`) | NPC dialogue, low-latency intent |
+| `hints-deep` | MLX-LM 32B Q4 (`:8500`) | Hint generation, longer reasoning |
+| `npc-ollama-7b` | Ollama 7B (`:11434`) | Legacy A/B comparator for `npc-fast` |
+| `hints-ollama-72b` | Ollama 72B (`:11434`) | Legacy A/B comparator for `hints-deep` |
+| `whisper-fr` / `stt-fr` | whisper.cpp (`:8300`) | French ASR |
+| `tts-zacus-primary` | F5-TTS-MLX via voice-bridge (`:8200`) | Zacus voice synthesis |
+
+## Persistence (headless macOS 26 — no launchd)
 
 Studio runs headless; `launchctl` over SSH is broken. All daemons run via
-`nohup` for current uptime + `@reboot` crontab for reboot persistence.
+`nohup` for current uptime + `@reboot` crontab for reboot persistence. Plus a
+`*/2` watchdog entry that relaunches voice-bridge on crash.
 
 ```cron
+@reboot /opt/homebrew/bin/node_exporter >> /Users/clems/node_exporter.log 2>&1
 @reboot /Users/clems/.local/bin/ollama serve >> /Users/clems/ollama.log 2>&1
 @reboot /Users/clems/whisper-install/whisper.cpp/build/bin/whisper-server -m /Users/clems/whisper-install/whisper.cpp/models/ggml-large-v3-turbo-q5_0.bin --port 8300 -t 8 --host 127.0.0.1 -l fr >> /Users/clems/whisper-server.log 2>&1
 @reboot /Users/clems/.local/bin/litellm --config /Users/clems/litellm-config/config.yaml --port 4000 --host 0.0.0.0 >> /Users/clems/litellm.log 2>&1
+@reboot /Users/clems/.local/bin/mlx_lm.server --model <hints-32b-q4> --port 8500 >> /Users/clems/mlx-32b.log 2>&1
+@reboot /Users/clems/.local/bin/mlx_lm.server --model <npc-7b-q4>   --port 8501 >> /Users/clems/mlx-7b.log  2>&1
+@reboot /Users/clems/voice-bridge/.venv/bin/uvicorn main:app --host 0.0.0.0 --port 8200 >> /Users/clems/voice-bridge.log 2>&1
+*/2 * * * * /Users/clems/voice-bridge/watchdog.sh >> /Users/clems/voice-bridge.watchdog.log 2>&1
 ```
 
 Edit on studio with `crontab -e`.
 
+## Resilience
+
+- **Watchdog** — `*/2 * * * * /Users/clems/voice-bridge/watchdog.sh` relance
+  voice-bridge si `pgrep -f "uvicorn main:app"` est vide (auto-recovery sous 2
+  minutes après un crash).
+- **Cache TTS disque** — `~/voice-bridge/cache/` indexé par
+  `sha256(text + voice_ref + steps)`. Hit cache = retour WAV immédiat (pas de
+  ré-synthèse). Purge :
+  ```bash
+  curl -X DELETE http://studio:8200/tts/cache
+  ```
+- **Graceful degradation** — `service_down.wav` (hardcoded fallback) servi par
+  voice-bridge si F5-TTS lève une exception irrécupérable. Le runtime ESP32
+  reçoit toujours un WAV valide, jamais un 5xx silencieux.
+
 ## Sync local → remote
 
 After editing here, push to studio:
@@ -41,21 +92,40 @@ ssh studio 'pkill -HUP -f "litellm --config" || pgrep -fl litellm'   # reload Li
 
 ```bash
 ssh studio 'curl -s http://localhost:11434/api/tags | jq .models[].name'   # ollama models
+ssh studio 'curl -s http://localhost:8500/v1/models'                       # mlx 32B (hints-deep)
+ssh studio 'curl -s http://localhost:8501/v1/models'                       # mlx 7B  (npc-fast)
 ssh studio 'curl -s http://localhost:4000/health/readiness'                # litellm
+ssh studio 'curl -s http://localhost:8200/healthz'                         # voice-bridge
 ssh studio 'curl -s -F "file=@/Users/clems/whisper-install/whisper.cpp/samples/jfk.wav" http://localhost:8300/inference'  # whisper
-ssh studio 'pgrep -fl ollama; pgrep -fl whisper-server; pgrep -fl litellm'  # daemons
+ssh studio 'pgrep -fl ollama; pgrep -fl whisper-server; pgrep -fl litellm; pgrep -fl mlx_lm.server; pgrep -fl "uvicorn main:app"'  # daemons
 ```
 
-## Known limitations (2026-05-03)
+## Pré-render NPC pool (F5-TTS)
 
-- `qwen2.5:72b-instruct-q8_0` via Ollama gives ~2 tok/s on M3 Ultra. Switch
-  to MLX-LM backend or Q4_K_M quant if hints latency unacceptable (cible
-  spec: <8 s for 50 tokens, currently ~25 s).
-- F5-TTS uses base multilingual checkpoint (no FR-specific). Quality
-  acceptable but no Zacus voice clone yet — needs `zacus_reference.wav` to
-  be recorded (P2).
-- Voice-bridge scaffold is **not running** as daemon yet. Activate when
-  F5-TTS in-process integration is wired.
+Generate the full NPC phrase pool against the live voice-bridge (idempotent —
+skips already-cached entries). Output: `hotline_tts/f5/<key>.wav` plus
+`hotline_tts/f5/manifest.json`.
+
+```bash
+python3 tools/tts/generate_npc_pool.py \
+  --backend f5 \
+  --voice-bridge-url http://studio:8200
+```
+
+Cache key derivation must mirror `voice-bridge/main.py::_voice_ref_token`. See
+`tools/CLAUDE.md` for backend semantics.
+
+## Known issues / TODO
+
+- **`steps` non borné** côté payload `/tts` — un client malicieux peut envoyer
+  `steps=999` et saturer la GPU queue. Ajouter `steps: int = Field(ge=1, le=64)`
+  dans le schéma Pydantic de voice-bridge.
+- **Default `--voice-bridge-url 192.168.0.150`** dans certains scripts — IP LAN
+  studio à confirmer selon le réseau du client. En itinérance préférer le
+  Tailscale `100.116.92.12` (ou alias DNS `studio`).
+- **`zacus_reference.wav`** à enregistrer (action humaine, prereq P2). Sans ce
+  fichier, F5-TTS-MLX tombe sur le checkpoint multilingue de base — qualité
+  acceptable mais pas de Zacus voice clone.
 
 ## Anti-patterns