docs: rewrite 9 instrument manuals in proper markdown format

All manuals follow openDAW convention: numbered sections, paragraph descriptions,
MIDI behavior documented, backend info, usage examples.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

packages/app/studio/public/manuals/devices/instruments/ace-step.md

@@ -1,36 +1,67 @@
 # IA AceStep
 
-Generateur de musique IA par prompt textuel. Produit des fichiers audio WAV.
+An AI music generator that creates audio from text prompts.
 
-**Type: Audio** — les samples generes se placent sur la timeline comme des regions audio.
+**Track type: Audio** — generated samples appear as regions on the timeline.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Release | Relachement du sample |
-| Tempo | BPM pour la generation (40-300) |
-| Duration | Duree du morceau genere (5-180s) |
+_IA AceStep_ generates music from text descriptions using the ACE-Step or MusicGen AI backend. The generated WAV is loaded as a playable sample. Use the AI panel to describe the music you want.
 
-## Panel IA
+Example uses:
 
-- **Prompt** : description textuelle de la musique souhaitee
-- **Style** : ambient, electronic, experimental, techno, house, drum-n-bass, jazz, classical, cinematic, etc.
-- **Composer** : lance la generation via AI Bridge (ACE-Step / MusicGen)
+- Background music for scenes
+- Ambient soundscapes from text
+- Quick musical sketches
+- AI-assisted composition
 
-## Exemples de prompts
+---
 
+## 1. Controls
+
+### 1.1 Volume
+
+Output level in dB.
+
+### 1.2 Release
+
+Sample release time.
+
+### 1.3 Tempo
+
+BPM hint for the generation (40-300).
+
+### 1.4 Duration
+
+Length of generated audio in seconds (5-180).
+
+---
+
+## 2. AI Panel
+
+### 2.1 Prompt
+
+Text description of the desired music. Examples:
 - "dark ambient with metallic textures and slow drones"
 - "upbeat electronic with arpeggiated synths"
 - "cinematic orchestral tension building"
-- "lo-fi hip hop chill beats"
 
-## Backend
+### 2.2 Style
 
-Utilise ACE-Step (si disponible) ou MusicGen via l'AI Bridge (port 8301).
-La generation peut prendre 20-60 secondes selon la duree.
+Select from 14 presets: ambient, electronic, experimental, classical, jazz, industrial, drone, glitch, techno, house, drum-n-bass, lo-fi, cinematic, world.
 
-*Instrument IA openDIAW.be*
+### 2.3 Composer
+
+Click to start generation. A progress bar shows the status. Generation takes 20-60 seconds depending on duration.
+
+---
+
+## 3. Backend
+
+Uses ACE-Step (port 9200) when available, falls back to MusicGen (port 9000) via the AI Bridge (port 8301).
+
+---
+
+_IA instrument — openDIAW.be / KXKM_

packages/app/studio/public/manuals/devices/instruments/circus.md

@@ -1,35 +1,66 @@
 # IA Circus
 
-Orgue de barbarie / calliope: synthese additive par tuyaux avec 4 registres harmoniques.
+A barrel organ / calliope synthesizer with pipe harmonics and mechanical wobble.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Register | Registre: Flute 8', Flute 4', Principal, Mixture |
-| Harmonics | Intensite des harmoniques superieures |
-| Tremolo Rate | Vitesse du tremolo (Hz) |
-| Tremolo Depth | Profondeur du tremolo |
-| Wobble | Instabilite de la hauteur (desaccord mecanique) |
-| Attack | Temps d'attaque du tuyau |
-| Release | Temps de fermeture du tuyau |
-| Air | Quantite de bruit d'air/souffle |
-| Detune | Desaccord entre tuyaux (cents) |
+_IA Circus_ simulates a steam-powered barrel organ (orgue de barbarie). Additive synthesis generates organ pipe tones with configurable harmonic content. Tremolo and wobble add the characteristic mechanical instability.
 
-## Registres
+Example uses:
 
-- **Flute 8'** : son doux, peu d'harmoniques
-- **Flute 4'** : une octave plus haut, plus clair
-- **Principal** : son plein avec harmoniques riches
-- **Mixture** : tous les harmoniques, son brillant et complexe
+- Fairground / circus atmospheres
+- Mechanical organ melodies
+- Vintage carousel music
+- Theatrical clown performance
 
-## Utilisation
+---
 
-Jouez des accords ou des melodies pour un son d'orgue de barbarie authentique.
-Le wobble simule l'instabilite mecanique d'un vrai orgue a manivelle.
-L'air ajoute le souffle caracteristique des tuyaux.
+## 1. Controls
 
-*Instrument IA openDIAW.be — spectacle de clown KXKM*
+### 1.1 Volume
+
+Output level in dB.
+
+### 1.2 Register
+
+Pipe register selection:
+- **Flute 8'** — soft, few harmonics
+- **Flute 4'** — one octave higher, brighter
+- **Principal** — full, rich harmonics
+- **Mixture** — all harmonics, brilliant and complex
+
+### 1.3 Harmonics
+
+Intensity of upper harmonics (0-100%).
+
+### 1.4 Tremolo Rate / Depth
+
+Speed and depth of amplitude tremolo.
+
+### 1.5 Wobble
+
+Mechanical pitch instability — simulates an imperfect hand-cranked mechanism.
+
+### 1.6 Attack / Release
+
+Pipe opening and closing time.
+
+### 1.7 Air
+
+Amount of breath/air noise mixed into the signal.
+
+### 1.8 Detune
+
+Slight detuning between pipes in cents.
+
+---
+
+## 2. MIDI
+
+Fully chromatic — play chords and melodies. Each note triggers an independent pipe voice.
+
+---
+
+_IA instrument — openDIAW.be / KXKM clown performance_

packages/app/studio/public/manuals/devices/instruments/drone.md

@@ -1,29 +1,74 @@
 # IA Drone
 
-Synthetiseur de pads et drones avec oscillateurs unison, filtre passe-bas et modulation LFO.
+A pad/drone synthesizer with unison oscillators, low-pass filter and LFO modulation.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Waveform | Forme d'onde: Sine, Triangle, Saw, Square |
-| Detune | Desaccord entre voix unison (cents) |
-| Unison | Nombre de voix (1, 3, 5, 7) |
-| Cutoff | Frequence du filtre passe-bas (Hz) |
-| Resonance | Resonance du filtre |
-| Attack | Temps d'attaque de l'enveloppe (s) |
-| Release | Temps de relachement (s) |
-| LFO Rate | Vitesse du LFO (Hz) |
-| LFO Depth | Profondeur de modulation LFO |
-| Stereo | Largeur stereo des voix unison |
+_IA Drone_ generates thick, evolving pads and drones. Multiple detuned oscillator voices create a wide stereo field. A low-pass filter shapes the tone, and an LFO adds movement.
 
-## Utilisation
+Example uses:
 
-Jouez des notes MIDI pour declencher des drones. Les voix unison creent de l'epaisseur.
-Augmentez le detune et le nombre de voix pour un son plus large.
-Le LFO module le volume pour des effets de pulsation lente.
+- Ambient pads and atmospheres
+- Evolving textured backgrounds
+- Bass drones for cinematic or live performance
+- Meditative soundscapes
 
-*Instrument IA openDIAW.be — performance live KXKM*
+---
+
+## 1. Controls
+
+### 1.1 Volume
+
+Output level in dB.
+
+### 1.2 Waveform
+
+Oscillator shape: Sine, Triangle, Sawtooth, or Square.
+
+### 1.3 Detune
+
+Spread between unison voices in cents. Higher values create a wider, chorus-like effect.
+
+### 1.4 Unison
+
+Number of simultaneous voices: 1, 3, 5 or 7. More voices = thicker sound.
+
+### 1.5 Cutoff
+
+Low-pass filter frequency in Hz. Lower values produce darker tones.
+
+### 1.6 Resonance
+
+Filter resonance (Q). Higher values emphasize frequencies around the cutoff.
+
+### 1.7 Attack
+
+Envelope attack time in seconds. Use long values (2-10s) for slow swells.
+
+### 1.8 Release
+
+Envelope release time after note-off. Long values (5-30s) for fading drones.
+
+### 1.9 LFO Rate
+
+Speed of the LFO modulation in Hz.
+
+### 1.10 LFO Depth
+
+Amount of LFO modulation applied to volume.
+
+### 1.11 Stereo
+
+Stereo width of the unison voices. 0% = mono center, 100% = full stereo spread.
+
+---
+
+## 2. MIDI
+
+Play any MIDI note to trigger a drone at that pitch. Multiple notes create polyphonic layers. Note-off triggers the release envelope.
+
+---
+
+_IA instrument — openDIAW.be / KXKM_

packages/app/studio/public/manuals/devices/instruments/glitch.md

@@ -1,26 +1,66 @@
 # IA Glitch
 
-Generateur de textures glitch: buffer repeat, stutter, bit crush et downsample.
+A glitch texture generator with buffer repeat, stutter, bit crush and downsample.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Buffer Size | Taille du buffer de capture (s) |
-| Repeat Rate | Frequence de repetition du buffer (Hz) |
-| Stutter | Probabilite de stutter (0-100%) |
-| Crush Bits | Resolution de bit crush (1-16 bits) |
-| Downsample | Facteur de sous-echantillonnage |
-| Pitch Shift | Decalage de hauteur (cents) |
-| Reverse | Probabilite de lecture inversee |
-| Dry/Wet | Balance signal original/traite |
+_IA Glitch_ captures audio into a buffer and applies destructive effects: stuttering repeats, bit crushing, downsampling, pitch shifting, and random reversal. Designed for chaotic textures and transitions.
 
-## Utilisation
+Example uses:
 
-Maintenez une note MIDI pour activer le glitch. Le buffer capture le bruit interne
-et le repete/ecrase/inverse selon les parametres. Ideal pour les transitions et les textures chaotiques.
+- Glitch transitions between sections
+- Noise textures for industrial/experimental music
+- Live performance chaos effects
+- Sound destruction and mangling
 
-*Instrument IA openDIAW.be*
+---
+
+## 1. Controls
+
+### 1.1 Volume
+
+Output level in dB.
+
+### 1.2 Buffer Size
+
+Size of the capture buffer in seconds (10ms to 1s).
+
+### 1.3 Repeat Rate
+
+How often the buffer repeats in Hz.
+
+### 1.4 Stutter
+
+Probability (0-100%) that a repeat cycle triggers stutter (rapid micro-repeats).
+
+### 1.5 Crush Bits
+
+Bit depth reduction. 16 = clean, 1 = extreme distortion.
+
+### 1.6 Downsample
+
+Sample rate reduction factor. 1x = clean, 64x = extreme lo-fi.
+
+### 1.7 Pitch Shift
+
+Pitch offset in cents applied to repeated buffer.
+
+### 1.8 Reverse
+
+Probability (0-100%) that a repeat plays backwards.
+
+### 1.9 Dry/Wet
+
+Balance between clean and processed signal.
+
+---
+
+## 2. MIDI
+
+Note-on activates the glitch engine. Note-off stops it. The glitch runs continuously while a note is held.
+
+---
+
+_IA instrument — openDIAW.be / KXKM_

packages/app/studio/public/manuals/devices/instruments/grain.md

@@ -1,27 +1,68 @@
 # IA Grain
 
-Synthetiseur granulaire: decoupe un sample en micro-grains et les rearrange aleatoirement.
+A granular synthesizer that chops a loaded sample into tiny grains and scatters them.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Position | Position de lecture dans le sample (0-100%) |
-| Grain Size | Taille des grains (5-500ms) |
-| Density | Nombre de grains par seconde (Hz) |
-| Pitch Spread | Variation aleatoire du pitch (cents) |
-| Position Spread | Variation aleatoire de la position |
-| Attack | Enveloppe d'attaque |
-| Release | Enveloppe de relachement |
-| Stereo | Dispersion stereo des grains |
+_IA Grain_ takes a loaded audio file and slices it into micro-grains (5-500ms). These grains are played back at random positions and pitches, creating evolving textures from any source material.
 
-## Utilisation
+Example uses:
 
-Chargez un sample via la zone de drop. Jouez une note MIDI pour declencher la granulation.
-La position controle ou dans le sample les grains sont extraits.
-Augmentez le pitch spread pour des textures etheriques.
+- Transforming vocals into ethereal textures
+- Creating ambient pads from field recordings
+- Glitchy rhythmic patterns from drum loops
+- Sound design and experimentation
 
-*Instrument IA openDIAW.be*
+---
+
+## 1. Controls
+
+### 1.1 Volume
+
+Output level in dB.
+
+### 1.2 Position
+
+Playback position in the loaded sample (0-100%).
+
+### 1.3 Grain Size
+
+Duration of each grain in seconds (5ms to 500ms).
+
+### 1.4 Density
+
+Number of grains triggered per second (1-100 Hz).
+
+### 1.5 Pitch Spread
+
+Random pitch variation applied to each grain in cents.
+
+### 1.6 Position Spread
+
+Random position offset for each grain. Higher values scatter grains across the sample.
+
+### 1.7 Attack / Release
+
+Envelope for note-on and note-off.
+
+### 1.8 Stereo
+
+Random stereo panning applied to each grain.
+
+---
+
+## 2. Sample Zone
+
+Click the {icon:Waveform} zone to load a sample. Drag & drop also works. The sample provides source material for granulation.
+
+---
+
+## 3. MIDI
+
+Each MIDI note triggers granular playback at a pitch ratio relative to C3 (middle C = original pitch).
+
+---
+
+_IA instrument — openDIAW.be / KXKM_

packages/app/studio/public/manuals/devices/instruments/honk.md

@@ -1,27 +1,64 @@
 # IA Honk
 
-Klaxon, sirene et corne parametrique avec sweep de frequence.
+A parametric klaxon, siren and horn with frequency sweep.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Mode | Klaxon (alternance 2 tons), Siren (sweep continu), Horn (ton stable + vibrato) |
-| Frequency | Frequence de base (Hz) |
-| Sweep Range | Amplitude du sweep (cents) |
-| Sweep Time | Duree d'un cycle de sweep (s) |
-| Richness | Intensite des harmoniques |
-| Attack | Temps d'attaque |
-| Release | Temps de relachement |
-| Noise | Quantite de bruit ajoute |
+_IA Honk_ generates klaxons, sirens and horns with configurable frequency sweep patterns. Three modes offer different sweep behaviors. The instrument is pitch-aware: MIDI notes shift the base frequency.
 
-## Modes
+Example uses:
 
-- **Klaxon** : alterne entre 2 frequences (classique pouet-pouet)
-- **Siren** : sweep sinusoidal continu haut-bas
-- **Horn** : ton stable avec leger vibrato (corne de brume)
+- Clown klaxon (pouet-pouet)
+- Emergency siren effects
+- Foghorn / ship horn
+- Comedic sound effects
 
-*Instrument IA openDIAW.be — spectacle de clown KXKM*
+---
+
+## 1. Controls
+
+### 1.1 Volume
+
+Output level in dB.
+
+### 1.2 Mode
+
+- **Klaxon** — alternates between two tones (classic honk)
+- **Siren** — continuous sinusoidal sweep up and down
+- **Horn** — steady tone with slight vibrato (foghorn)
+
+### 1.3 Frequency
+
+Base frequency in Hz. MIDI notes shift this — C3 = base, higher notes = higher pitch.
+
+### 1.4 Sweep Range
+
+Amplitude of the frequency sweep in cents.
+
+### 1.5 Sweep Time
+
+Duration of one sweep cycle in seconds.
+
+### 1.6 Richness
+
+Harmonic content intensity.
+
+### 1.7 Attack / Release
+
+Onset and decay time.
+
+### 1.8 Noise
+
+Amount of broadband noise mixed in.
+
+---
+
+## 2. MIDI
+
+Note-on triggers the honk. MIDI pitch shifts the base frequency. Note-off starts the release.
+
+---
+
+_IA instrument — openDIAW.be / KXKM clown performance_

packages/app/studio/public/manuals/devices/instruments/kokoro-tts.md

@@ -1,44 +1,75 @@
 # IA Kokoro TTS
 
-Synthese vocale rapide avec 12 voix. Produit des fichiers audio WAV.
+Fast text-to-speech synthesis with 12 voices.
 
-**Type: Audio** — les voix generees se placent sur la timeline.
+**Track type: Audio** — generated speech appears as a region on the timeline.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Release | Relachement du sample |
-| Speed | Vitesse de parole (0.5x - 2.0x) |
+_IA Kokoro_ synthesizes speech from text using the Kokoro TTS engine. Choose from 12 voices across 3 languages (English, French, British English). Generated WAV files are loaded as playable samples.
 
-## Panel IA
+Example uses:
 
-- **Texte** : le texte a prononcer (textarea multi-lignes)
-- **Voix** : selection parmi 12 voix Kokoro
-- **Synthetiser** : lance la generation via Kokoro TTS
+- Voiceover and narration
+- Character dialogue for scenes
+- Spoken word over music
+- Multilingual announcements
 
-## Voix disponibles
+---
 
-| ID | Description |
-|---|---|
-| ff_siwis | Siwis (francais, feminin) |
-| af_heart | Heart (anglais, feminin) |
-| af_bella | Bella (anglais, feminin) |
-| af_nicole | Nicole (anglais, feminin) |
-| af_sarah | Sarah (anglais, feminin) |
-| af_sky | Sky (anglais, feminin) |
-| am_adam | Adam (anglais, masculin) |
-| am_michael | Michael (anglais, masculin) |
-| bf_emma | Emma (britannique, feminin) |
-| bf_isabella | Isabella (britannique, feminin) |
-| bm_george | George (britannique, masculin) |
-| bm_lewis | Lewis (britannique, masculin) |
+## 1. Controls
 
-## Backend
+### 1.1 Volume
 
-Kokoro TTS (port 9201) — generation rapide, ~2-5 secondes.
+Output level in dB.
 
-*Instrument IA openDIAW.be*
+### 1.2 Release
+
+Sample release time.
+
+### 1.3 Speed
+
+Playback speed multiplier (0.5x to 2.0x).
+
+---
+
+## 2. AI Panel
+
+### 2.1 Text
+
+Multi-line text input. Enter the text to be spoken.
+
+### 2.2 Voice
+
+Select from 12 voices:
+
+| Voice | Language | Gender |
+|---|---|---|
+| ff_siwis | French | Female |
+| af_heart | English | Female |
+| af_bella | English | Female |
+| af_nicole | English | Female |
+| af_sarah | English | Female |
+| af_sky | English | Female |
+| am_adam | English | Male |
+| am_michael | English | Male |
+| bf_emma | British | Female |
+| bf_isabella | British | Female |
+| bm_george | British | Male |
+| bm_lewis | British | Male |
+
+### 2.3 Synthesize
+
+Click to generate. A progress bar shows the status. Generation takes 2-5 seconds.
+
+---
+
+## 3. Backend
+
+Kokoro TTS engine (port 9201) via the AI Bridge.
+
+---
+
+_IA instrument — openDIAW.be / KXKM_

packages/app/studio/public/manuals/devices/instruments/magenta.md

@@ -1,35 +1,74 @@
 # IA Magenta
 
-Generateur MIDI algorithmique avec synthese polyphonique integree.
+An algorithmic MIDI sequence generator with built-in polyphonic synthesizer.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Waveform | Sine, Triangle, Saw, Square |
-| Attack | Enveloppe d'attaque |
-| Release | Enveloppe de relachement |
-| Cutoff | Filtre passe-bas |
-| Model | Melody (gammes), Drums (percussions), Improv (chromatique) |
-| Temperature | Aleatoire: bas=previsible, haut=chaotique |
-| Steps | Nombre de pas de la sequence (8-128) |
-| Density | Probabilite qu'un pas contienne une note |
-| Stereo | Dispersion stereo par hauteur de note |
+_IA Magenta_ generates MIDI note sequences using algorithmic composition and plays them through an integrated synthesizer. Three models offer different musical styles. No external connection required — all generation runs locally.
 
-## Gammes disponibles
+Example uses:
 
-- Melody: mineur naturel
-- Drums: kick, snare, hihat, ride, crash, tom
-- Improv: pentatonique avec variations
+- Generative melodies and arpeggios
+- Algorithmic drum patterns
+- Improvisational sequences
+- Live performance with controlled randomness
 
-## Utilisation
+---
 
-1. Reglez Model, Temperature et Steps
-2. Cliquez **Generate** pour creer une sequence
-3. Cliquez **Play** pour ecouter
-4. Les notes MIDI jouees declenchent aussi le synth integre
+## 1. Controls
 
-*Instrument IA openDIAW.be*
+### 1.1 Volume
+
+Output level in dB.
+
+### 1.2 Waveform
+
+Synth oscillator: Sine, Triangle, Saw, Square.
+
+### 1.3 Attack / Release
+
+Synth amplitude envelope.
+
+### 1.4 Cutoff
+
+Low-pass filter frequency.
+
+### 1.5 Model
+
+Generation algorithm:
+- **Melody** — minor scale patterns
+- **Drums** — kick, snare, hihat, ride, crash, tom
+- **Improv** — pentatonic with chromatic variations
+
+### 1.6 Temperature
+
+Randomness: low = predictable patterns, high = chaotic.
+
+### 1.7 Steps
+
+Sequence length (8 to 128 steps).
+
+### 1.8 Density
+
+Probability that each step contains a note (0-100%).
+
+### 1.9 Stereo
+
+Stereo panning spread based on note pitch.
+
+---
+
+## 2. Generation
+
+1. Set Model, Temperature, Steps and Density
+2. Click **Generate** to create a sequence
+3. Click **Play** to listen
+4. Click **Stop** to halt playback
+
+MIDI notes also trigger the built-in synth directly.
+
+---
+
+_IA instrument — openDIAW.be / KXKM_

packages/app/studio/public/manuals/devices/instruments/piper.md

@@ -1,36 +1,65 @@
 # IA Piper
 
-Synthese vocale Piper/Chatterbox avec voix de personas. Produit des fichiers audio WAV.
+Text-to-speech synthesis with persona voices (Piper/Chatterbox).
 
-**Type: Audio** — les voix generees se placent sur la timeline.
+**Track type: Audio** — generated speech appears as a region on the timeline.
 
 ---
 
-## Controles
+## 0. Overview
 
-| Parametre | Description |
-|---|---|
-| Volume | Niveau de sortie (dB) |
-| Release | Relachement du sample |
-| Speed | Vitesse de parole (0.5x - 2.0x) |
+_IA Piper_ synthesizes speech using persona-specific voices. Each persona from the 3615-KXKM system has a mapped French voice. Chatterbox provides high-quality zero-shot voice cloning as a backend.
 
-## Panel IA
+Example uses:
 
-- **Texte** : le texte a prononcer
-- **Persona** : voix du personnage (pharmacius, schaeffer, merzbow, cage, radigue, sunra, haraway, batty, deleuze, turing)
-- **Prononcer** : lance la generation via Piper/Chatterbox
+- Character voices for performances
+- Persona narration
+- French text-to-speech
+- Voice acting placeholders
 
-## Personas et voix
+---
 
-Chaque persona est mappee a une voix Piper francaise:
-- **Feminin chaud** (siwis-medium): Radigue, Oliveros, Haraway, Curie
-- **Masculin academique** (upmc-medium): Pharmacius, Foucault, Deleuze, Turing
-- **Masculin naturel** (tom-medium): Schaeffer, Cage, Tarkovski, Picasso
-- **Masculin brut** (gilles-low): Sun Ra, Merzbow, Swartz, Grotowski
+## 1. Controls
 
-## Backend
+### 1.1 Volume
 
-Piper TTS / Chatterbox (port 9100) — generation ~3-8 secondes.
-Chatterbox offre une qualite superieure avec clonage vocal zero-shot.
+Output level in dB.
 
-*Instrument IA openDIAW.be*
+### 1.2 Release
+
+Sample release time.
+
+### 1.3 Speed
+
+Playback speed multiplier (0.5x to 2.0x).
+
+---
+
+## 2. AI Panel
+
+### 2.1 Text
+
+Multi-line text input. Enter the text to be spoken.
+
+### 2.2 Persona
+
+Select the speaking character. Each persona maps to a French Piper voice:
+
+- **Female warm** (siwis): Radigue, Oliveros, Haraway, Curie
+- **Male academic** (upmc): Pharmacius, Foucault, Deleuze, Turing
+- **Male natural** (tom): Schaeffer, Cage, Tarkovski, Picasso
+- **Male rough** (gilles): Sun Ra, Merzbow, Swartz, Grotowski
+
+### 2.3 Pronounce
+
+Click to generate. Progress bar shows status. Generation takes 3-8 seconds.
+
+---
+
+## 3. Backend
+
+Piper TTS / Chatterbox (port 9100) via the AI Bridge. Chatterbox offers superior quality with zero-shot voice cloning.
+
+---
+
+_IA instrument — openDIAW.be / KXKM_