Wiself/Voice

1# voice — Help2 3Generated from `voice --help` and `voice <cmd> --help` (v3.1, 14 commands).4 5## voice --help6 7```text8usage: voice [-h] [-v] [--non-interactive] [--version]9             [--completion {bash,zsh,fish}]10             {get,cast,delta,graft,chorus,add,list,info,remove,cleanup,auth,report,path,doctor} ...11 12Lift a model's voice and cast it onto another model.13 14positional arguments:15  {get,cast,delta,graft,chorus,add,list,info,remove,cleanup,auth,report,path,doctor}16    get                 Get a voice from a model (safetensors repo, .gguf, or17                        URL)18    cast                Cast a voice onto a .gguf or .safetensors file19    delta               Make a reusable style (voice − base)20    graft               Graft one voice onto another (DELLA single graft)21    chorus              Blend voices: pulse + support + lead (DELLA chorus)22    add                 Add a voice from local voice.safetensors + voice.json23    list                List stored voices24    info                Show voice details25    remove              Remove a stored voice26    cleanup             Remove incomplete downloads27    auth                Save your HF token (optional)28    report              Save a report locally and optionally share to29                        community30    path                Put voice on your PATH31    doctor              Check your setup32 33options:34  -h, --help            show this help message and exit35  -v, --verbose         Verbose: tensor shapes, decisions, and paths36  --non-interactive     Fail instead of prompting (same as piped stdin; needs37                        --yes to overwrite/remove)38  --version             Show version and environment39  --completion {bash,zsh,fish}40                        Print shell completion script41```42 43## voice get44 45```text46usage: voice get [-h] [-q] [--name NAME] [--multi] [--out OUT] [--yes]47                 source [tensors ...]48 49positional arguments:50  source       HF model ID, .gguf/.safetensors path, or URL51  tensors      Tensor name(s) to fetch (default: auto-detect output head; 2+52               or --multi saves a voicepack)53 54options:55  -h, --help   show this help message and exit56  -q, --quiet  Quieter: progress and success notes off (warnings and errors57               still print)58  --name NAME  Voice name to save (registry voices only; not with --out)59  --multi      Fetch the named tensors (1+) as voicepack.safetensors +60               voicepack.json (not with --name)61  --out OUT    Write a standalone file instead of a registry voice (default62               pack name: voicepack.safetensors)63  --yes        Assume yes to prompts (overwrite without asking)64```65 66## voice cast67 68```text69usage: voice cast [-h] [--out OUT] [--yes] [--quant QUANT]70                  [--exclude EXCLUDES] [--speak] [-q]71                  target voice [tensors ...]72 73positional arguments:74  target              Path to the target .gguf or .safetensors75  voice               Voice name, or a .safetensors/.gguf path76  tensors             Only cast these tensors (default: every tensor in the77                      voice)78 79options:80  -h, --help          show this help message and exit81  --out OUT           Output file or folder (default:82                      ~/Voiced/<voice>-<model>/<voice>-<model>.<ext matching83                      target>; a folder invents the filename)84  --yes               Assume yes to prompts (overwrite without asking)85  --quant QUANT       Output quant override: keep (default, preserve quality)86                      | target | any quant87                      (q8_0,q4_0,q4_k,q6_k,iq4_nl,f16,f32,bf16,...;88                      safetensors targets: keep|target|f32|f16|bf16)89  --exclude EXCLUDES  Skip tensors on cast (repeatable: --exclude foo90                      --exclude bar, or --exclude foo,bar)91  --speak             Voice speaks through output.weight (invented if absent92                      when head/embed geometry survives; GGUF targets only)93  -q, --quiet         Quieter: progress and success notes off (warnings and94                      errors still print)95```96 97## voice delta98 99```text100usage: voice delta [-h] [--base BASE] [--yes] [-q] [voicename]101 102positional arguments:103  voicename    Voice name (default: pick from registry)104 105options:106  -h, --help   show this help message and exit107  --base BASE  Base HF repo (org/model) or local path to base safetensors108  --yes        Assume yes to prompts (overwrite without asking)109  -q, --quiet  Quieter: progress and success notes off (warnings and errors110               still print)111```112 113## voice graft114 115```text116usage: voice graft [-h] [--density DENSITY] [--epsilon EPSILON]117                   [--lambda_ LAMBDA_] [--rescale {l1,inv_p}] [--seed SEED]118                   [--out OUT] [--yes] [-q]119                   [pulse] [donor]120 121positional arguments:122  pulse                 Pulse voice name (default: pick from registry)123  donor                 Donor voice name (default: pick from registry)124 125options:126  -h, --help            show this help message and exit127  --density DENSITY     Keep-fraction for magnitude pruning (default: 0.8)128  --epsilon EPSILON     Rank-ramp half-width (default: 0.1)129  --lambda_ LAMBDA_     Global rescale after combine (default: 1.0)130  --rescale {l1,inv_p}  Prune rescale: l1 (stable) or inv_p (paper-faithful)131  --seed SEED           Bernoulli seed for reproducible grafts (default: 42)132  --out OUT             Output voice name (default: chorus-<pulse>-<donor>)133  --yes                 Assume yes to prompts (overwrite without asking)134  -q, --quiet           Quieter: progress and success notes off (warnings and135                        errors still print)136```137 138## voice chorus139 140```text141usage: voice chorus [-h] [--config] [--density DENSITY]142                    [--lead-density LEAD_DENSITY] [--epsilon EPSILON]143                    [--lead-epsilon LEAD_EPSILON] [--lambda_ LAMBDA_]144                    [--rescale {l1,inv_p}] [--seed SEED] [--out OUT] [--yes]145                    [-q]146                    [pulse] [donors ...]147 148positional arguments:149  pulse                 Pulse voice name (default: pick interactively)150  donors                Support + lead voice names (0-2; 1 = duet at lead151                        density)152 153options:154  -h, --help            show this help message and exit155  --config              Stage knob screens per graft (Enter accepts each156                        default; screens prefill from flags)157  --density DENSITY     First-graft density (default: 0.8)158  --lead-density LEAD_DENSITY159                        Lead-graft density (default: 0.9)160  --epsilon EPSILON     Rank-ramp half-width (default: 0.1)161  --lead-epsilon LEAD_EPSILON162                        Lead-graft epsilon (default: 0.05; keeps 0.9±eps163                        valid)164  --lambda_ LAMBDA_     Global rescale after combine (default: 1.0)165  --rescale {l1,inv_p}  Prune rescale: l1 (stable) or inv_p (paper-faithful)166  --seed SEED           Bernoulli seed for reproducible choruses (default: 42)167  --out OUT             Final output voice name (default:168                        chorus-<pulse>-<lead>)169  --yes                 Assume yes to prompts (overwrite without asking)170  -q, --quiet           Quieter: progress and success notes off (warnings and171                        errors still print)172```173 174## voice add175 176```text177usage: voice add [-h] [--name NAME] [--yes] [-q] safetensors json178 179positional arguments:180  safetensors  Path to voice.safetensors (required)181  json         Path to voice.json (required)182 183options:184  -h, --help   show this help message and exit185  --name NAME  Voice name to save (default from voice.json)186  --yes        Auto-confirm overwrite187  -q, --quiet  Quieter: progress and success notes off (warnings and errors188               still print)189```190 191## voice list192 193```text194usage: voice list [-h] [--base] [-q] [base]195 196positional arguments:197  base         Show cached base heads (same as --base)198 199options:200  -h, --help   show this help message and exit201  --base       Show cached base heads instead of voices202  -q, --quiet  Quieter: progress and success notes off (warnings and errors203               still print)204```205 206## voice info207 208```text209usage: voice info [-h] [-q] voicename210 211positional arguments:212  voicename    Voice name213 214options:215  -h, --help   show this help message and exit216  -q, --quiet  Quieter: progress and success notes off (warnings and errors217               still print)218```219 220## voice remove221 222```text223usage: voice remove [-h] [--yes] [-q] voice224 225positional arguments:226  voice        Voice name227 228options:229  -h, --help   show this help message and exit230  --yes        Assume yes to prompts (remove without asking)231  -q, --quiet  Quieter: progress and success notes off (warnings and errors232               still print)233```234 235## voice cleanup236 237```text238usage: voice cleanup [-h] [--force] [-q]239 240options:241  -h, --help   show this help message and exit242  --force      Also clear in-progress downloads (locked)243  -q, --quiet  Quieter: progress and success notes off (warnings and errors244               still print)245```246 247## voice auth248 249```text250usage: voice auth [-h] [--clear] [--show] [-q]251 252options:253  -h, --help   show this help message and exit254  --clear      Remove the stored token255  --show       Show the masked token256  -q, --quiet  Quieter: progress and success notes off (warnings and errors257               still print)258```259 260## voice report261 262```text263usage: voice report [-h] [--message MESSAGE] [--voice VOICE] [--target TARGET]264                    [--reporter REPORTER] [--yes] [--no-push] [--sync] [-q]265 266options:267  -h, --help            show this help message and exit268  --message, -m MESSAGE269                        Freeform note (240ch, sanitized)270  --voice VOICE         Voice name context271  --target TARGET       Target file basename context272  --reporter REPORTER   Reporter id (optional, 64ch)273  --yes                 Non-interactive: skip preview and auto-share to274                        community275  --no-push             Local only, skip share prompt276  --sync                Sync pending outbox to community277  -q, --quiet           Quieter: progress and success notes off (warnings and278                        errors still print)279```280 281## voice path282 283```text284usage: voice path [-h] [-q]285 286options:287  -h, --help   show this help message and exit288  -q, --quiet  Quieter: progress and success notes off (warnings and errors289               still print)290```291 292## voice doctor293 294```text295usage: voice doctor [-h] [-q]296 297options:298  -h, --help   show this help message and exit299  -q, --quiet  Quieter: progress and success notes off (warnings and errors300               still print)301```302