@qvac/tts-ggml 0.13.0 → 0.14.1

package/index.d.ts

@@ -1,823 +1,9 @@
 import { type QvacResponse } from "@qvac/infer-base";
 import * as errorModule from "./lib/error";
-import { assessFit as assessFitImpl, type Audio8FitRequest, type MossFitRequest, type MossSoundEffectFitRequest, type ChatterboxFitRequest, type CosyvoiceFitRequest, type ParlerFitRequest, type SupertonicFitRequest, type TtsFitEngine, type TtsFitRequest, type TtsFitResult, type TtsFitStatus } from "./lib/fit";
-import { type SentenceDelimiterPreset } from "./lib/textStreamAccumulator";
-declare const ENGINE_CHATTERBOX = "chatterbox";
-declare const ENGINE_SUPERTONIC = "supertonic";
-declare const ENGINE_COSYVOICE3 = "cosyvoice3";
-declare const ENGINE_PARLER = "parler";
-declare const ENGINE_AUDIO8 = "audio8";
-declare const ENGINE_MOSS = "moss";
-declare const ENGINE_MOSS_SFX = "moss-sfx";
-declare const ENGINE_MOSS_SPEECH = "moss-speech";
-declare const ENGINE_POCKET = "pocket";
-declare const COSYVOICE_DIALECTS: {
-    readonly cantonese: "广东话";
-    readonly northeastern: "东北话";
-    readonly gansu: "甘肃话";
-    readonly guizhou: "贵州话";
-    readonly henan: "河南话";
-    readonly hubei: "湖北话";
-    readonly hunan: "湖南话";
-    readonly jiangxi: "江西话";
-    readonly minnan: "闽南话";
-    readonly ningxia: "宁夏话";
-    readonly shanxi: "山西话";
-    readonly shaanxi: "陕西话";
-    readonly shandong: "山东话";
-    readonly shanghai: "上海话";
-    readonly sichuan: "四川话";
-    readonly tianjin: "天津话";
-    readonly yunnan: "云南话";
-};
-declare const COSYVOICE_VOLUMES: {
-    readonly loud: "Please say a sentence as loudly as possible.";
-    readonly soft: "Please say a sentence in a very soft voice.";
-};
-declare const COSYVOICE_STYLES: {
-    readonly peppa: "我想体验一下小猪佩奇风格，可以吗？";
-    readonly robot: "你可以尝试用机器人的方式解答吗？";
-};
-/**
- * CosyVoice3 controls that have no canonical cross-engine vocabulary yet.
- * Emotion and speaking rate are NOT here -- use the top-level `emotion` /
- * `pace` options, which work the same way on every engine that supports them.
- * Exactly one field takes effect per synthesis, resolved by precedence
- * dialect > volume > style. Pass a raw string instead for an arbitrary
- * instruction (advanced escape hatch).
- */
-interface CosyvoiceInstruct {
-    /** Chinese dialect; renders "请用{dialect}表达。". */
-    dialect?: keyof typeof COSYVOICE_DIALECTS;
-    /** Loudness. */
-    volume?: keyof typeof COSYVOICE_VOLUMES;
-    /** Playful style preset. */
-    style?: keyof typeof COSYVOICE_STYLES;
-}
-declare const EMOTIONS: readonly ["command", "anger", "narration", "conversation", "disgust", "fear", "happy", "neutral", "proper noun", "news", "sad", "surprise"];
-declare const PACES: readonly ["slow", "moderate", "fast"];
-type Emotion = (typeof EMOTIONS)[number];
-type Pace = (typeof PACES)[number];
-declare const MOSS_SPEECH_ROLES: readonly ["system", "user", "assistant"];
-type EngineType = typeof ENGINE_CHATTERBOX | typeof ENGINE_SUPERTONIC | typeof ENGINE_COSYVOICE3 | typeof ENGINE_PARLER | typeof ENGINE_AUDIO8 | typeof ENGINE_MOSS | typeof ENGINE_MOSS_SFX | typeof ENGINE_MOSS_SPEECH | typeof ENGINE_POCKET;
-/**
- * Model file paths for the GGML TTS backend. Engine is auto-detected
- * from these fields (Chatterbox vs Supertonic) unless overridden via
- * `TTSGgmlOptions.engine`. All paths must be absolute and are passed
- * through to the native layer as-is.
- */
-interface TTSGgmlFiles {
-    pocketFlowModel?: string;
-    pocketMimiModel?: string;
-    pocketFrontend?: string;
-    pocketVoice?: string;
-    /**
-     * Bundle root. For Chatterbox, expected to contain
-     * `chatterbox-t3-turbo.gguf` + `chatterbox-s3gen.gguf` (turbo) or
-     * `chatterbox-t3-mtl.gguf` + `chatterbox-s3gen-mtl.gguf` (multilingual).
-     * For Supertonic, expected to contain `supertonic.gguf`.
-     */
-    modelDir?: string;
-    /** Chatterbox T3 (text to speech tokens) GGUF path. Overrides `modelDir`. */
-    t3Model?: string;
-    t3ModelPath?: string;
-    t3?: string;
-    /** Chatterbox S3Gen + HiFT (speech tokens to 24 kHz wav) GGUF path. Overrides `modelDir`. */
-    s3genModel?: string;
-    s3genModelPath?: string;
-    s3gen?: string;
-    /** Supertonic single-file GGUF path. Overrides `modelDir`. */
-    supertonicModel?: string;
-    supertonicModelPath?: string;
-    supertonic?: string;
-    /** Parler single-file GGUF path (mini/large/indic). Overrides `modelDir`. */
-    parlerModel?: string;
-    parlerModelPath?: string;
-    parler?: string;
-    /** Audio8 DualAR language model GGUF path. Overrides `modelDir`. */
-    audio8Lm?: string;
-    audio8LmPath?: string;
-    /** Audio8 codec synthesis half (codes to 44.1 kHz wav). Overrides `modelDir`. */
-    audio8CodecDecoder?: string;
-    audio8CodecDecoderPath?: string;
-    /**
-     * Audio8 codec analysis half (wav to codes). Only needed to clone a voice
-     * from a recording; a text-only deployment can leave it out.
-     */
-    audio8CodecEncoder?: string;
-    audio8CodecEncoderPath?: string;
-    /**
-     * MOSS Delay backbone GGUF path: MOSS-TTS (`moss-tts-delay-*.gguf`) or the
-     * MOSS-TTSD dialogue checkpoint (`moss-ttsd-*.gguf`). Overrides `modelDir`.
-     */
-    mossBackbone?: string;
-    mossBackbonePath?: string;
-    /** MOSS codec synthesis half (codes to 24 kHz wav). Overrides `modelDir`. */
-    mossCodecDecoder?: string;
-    mossCodecDecoderPath?: string;
-    /**
-     * MOSS codec analysis half (wav to codes). Only needed to clone a voice
-     * from `referenceAudio` or for `dialogueReferences`; a text-only deployment
-     * can leave it out.
-     */
-    mossCodecEncoder?: string;
-    mossCodecEncoderPath?: string;
-    /**
-     * MOSS-SoundEffect GGUF (`moss-sfx-*.gguf`): text encoder, DiT and VAE in
-     * one file. Routes to the `moss-sfx` engine. Overrides `modelDir`.
-     */
-    mossSoundEffect?: string;
-    mossSoundEffectPath?: string;
-    /**
-     * MOSS-Speech language model GGUF (`moss-speech-*.gguf`, e.g.
-     * `moss-speech-q8_0.gguf`). Routes to the `moss-speech` engine together
-     * with `mossSpeechCodec`. Overrides `modelDir`.
-     */
-    mossSpeechModel?: string;
-    mossSpeechModelPath?: string;
-    /**
-     * MOSS-Speech codec GGUF (`moss-speech-codec-*.gguf`): speech tokenizer,
-     * reply decoder and the default reply voice. Overrides `modelDir`.
-     */
-    mossSpeechCodec?: string;
-    mossSpeechCodecPath?: string;
-    /**
-     * CosyVoice3 model directory holding the sub-model GGUFs
-     * (`cosyvoice3-{llm,flow,hift}-*.gguf`) plus `voice.gguf`, `vocab.json` and
-     * `merges.txt`. Routes to the CosyVoice3 engine. Falls back to the shared
-     * `modelDir` when unset.
-     */
-    cosyvoiceModelDir?: string;
-    /** CosyVoice3 per-component GGUF paths (override discovery under the model dir). */
-    cosyvoiceLlmModel?: string;
-    cosyvoiceLlmModelPath?: string;
-    cosyvoiceFlowModel?: string;
-    cosyvoiceFlowModelPath?: string;
-    cosyvoiceHiftModel?: string;
-    cosyvoiceHiftModelPath?: string;
-    /**
-     * CosyVoice3 voice-cloning add-on GGUFs, required only when
-     * `referenceAudio` is set: the speech_tokenizer_v3 speech tokenizer
-     * (`cosyvoice3-s3tok-*.gguf`) and the CAM++ speaker encoder
-     * (`cosyvoice3-campplus-*.gguf`). Auto-discovered under
-     * `cosyvoiceModelDir` by those name prefixes when unset.
-     */
-    cosyvoiceS3tokModel?: string;
-    cosyvoiceS3tokModelPath?: string;
-    cosyvoiceCampplusModel?: string;
-    cosyvoiceCampplusModelPath?: string;
-    /**
-     * CosyVoice3 Qwen2 BPE text frontend (`vocab.json` + `merges.txt`) and the
-     * baked default-voice GGUF (`voice.gguf`). Each is resolved under the model
-     * dir when unset; `cosyvoiceVoiceModel` is how a different baked voice is
-     * selected without a separate model directory.
-     */
-    cosyvoiceVocab?: string;
-    cosyvoiceVocabPath?: string;
-    cosyvoiceMerges?: string;
-    cosyvoiceMergesPath?: string;
-    cosyvoiceVoiceModel?: string;
-    cosyvoiceVoiceModelPath?: string;
-    /**
-     * LavaSR enhancer GGUF: single-file Vocos bandwidth extension produced by
-     * tts-cpp/scripts/convert-lavasr-enhancer-to-gguf.py. When supplied, output
-     * is neurally upsampled to 48 kHz (the canonical way to enable enhancement;
-     * `enhancer.enhancerPath` is the only alternative).
-     */
-    lavasrEnhancer?: string;
-    /**
-     * LavaSR denoiser GGUF: UL-UNAS speech denoiser produced by
-     * tts-cpp/scripts/convert-lavasr-denoiser-to-gguf.py. Runs before the
-     * enhancer and is rate-preserving (the canonical way to enable denoising;
-     * `denoiser.denoiserPath` is the only alternative).
-     *
-     * The tts-cpp UL-UNAS forward is implemented in qvac-fabric-speech.cpp
-     * PR #78 (scalar CPU port, validated bit-close to the ONNX reference).
-     */
-    lavasrDenoiser?: string;
-    /** Optional directory containing baked Chatterbox voice profiles. */
-    voicesDir?: string;
-    /**
-     * Chatterbox MTL only: directory holding the compiled MeCab/IPAdic
-     * dictionary used for Japanese morphological segmentation. Forwarded to
-     * tts-cpp's `EngineOptions::mecab_dict_path`. Alias: top-level
-     * `mecabDictPath`.
-     */
-    mecabDictDir?: string;
-    mecabDictPath?: string;
-    /**
-     * Chatterbox MTL only: path to the Cangjie TSV used for Chinese
-     * romanisation. Forwarded to tts-cpp's `EngineOptions::cangjie_tsv_path`.
-     */
-    cangjieTsvPath?: string;
-    cangjieTsv?: string;
-}
-interface TTSGgmlRuntimeConfig {
-    /**
-     * Language code; default "en". Chatterbox MTL accepts
-     * es/fr/de/pt/it/zh/ja/ko/... CosyVoice3: reserved / not yet effective — the
-     * text-normalization frontend is not yet integrated, so it is accepted but
-     * not acted on.
-     */
-    language?: string;
-    /**
-     * Route inference through a GPU backend (Metal / Vulkan / OpenCL) if
-     * available. Defaults to `false`. Audio8 uses Vulkan on Linux and Windows;
-     * CosyVoice3 selects Metal on Apple, Vulkan on desktop Linux / Windows, and
-     * OpenCL/Adreno on Android (Mali / Xclipse decline to CPU); the other
-     * GPU-capable engines select a backend for the host platform.
-     */
-    useGPU?: boolean;
-    /**
-     * Desired output sample rate in Hz (8000-192000); omit to keep the engine's
-     * native rate. Resamples the native output (24 kHz for Chatterbox and
-     * CosyVoice3; 44.1 kHz for Supertonic, Parler, and Audio8), or the 48 kHz
-     * LavaSR-enhanced signal, before emitting. `TTSOutputChunk.sampleRate`
-     * reports the resulting rate.
-     *
-     * CosyVoice3 native chunk streaming emits at 24 kHz: a different rate is
-     * only accepted there when the LavaSR enhancer is active, because the
-     * enhancer's overlap-reprocess window resamples without chunk seams.
-     *
-     * MOSS emits its native 24 kHz only and rejects any other rate.
-     */
-    outputSampleRate?: number;
-    backendsDir?: string;
-    openclCacheDir?: string;
-    vulkanCacheDir?: string;
-}
-/**
- * LavaSR enhancer config. The discriminated `type` leaves room for future
- * enhancer kinds; v1 ships `lavasr`. Enhancement is enabled by providing a
- * GGUF path here or via `files.lavasrEnhancer`.
- */
-interface LavaSREnhancerOptions {
-    type: "lavasr";
-    /** Enhancer GGUF path (alternative to `files.lavasrEnhancer`). */
-    enhancerPath?: string;
-}
-/**
- * LavaSR denoiser config. V1 ships the `lavasr` UL-UNAS denoiser. Denoising
- * is enabled by providing a GGUF path here or via `files.lavasrDenoiser`.
- * It runs before the enhancer and preserves the sample rate.
- */
-interface LavaSRDenoiserOptions {
-    type: "lavasr";
-    /** Denoiser GGUF path (alternative to `files.lavasrDenoiser`). */
-    denoiserPath?: string;
-}
-/**
- * Cross-engine conditioning. Accepted at construction and on `reload()` by
- * every engine that supports them, and per call by Parler and CosyVoice3. A
- * value outside the canonical vocabulary, outside the engine's supported
- * subset, or on a channel the engine cannot change per call, throws naming the
- * alternative -- nothing is silently degraded.
- */
-interface TTSConditioningFields {
-    /** Speaking style. Parler: all 12. CosyVoice3: anger|happy|neutral|sad. */
-    emotion?: Emotion;
-    /**
-     * Speaking rate. Parler / CosyVoice3 / Supertonic. Supertonic conditions its
-     * engine at construction, so its pace only moves there or via `reload()`.
-     */
-    pace?: Pace;
-}
-/**
- * Parler voice-description inputs. Either a free-text `description` (alias
- * `voiceDescription`) or the template fields, rendered natively through
- * tts-cpp's build_description(); mixing the two at the same level is rejected.
- * Accepted at construction, on `reload()`, and per call (Parler only).
- * `emotion` and `pace` moved to TTSConditioningFields -- Parler still renders
- * both, but they are no longer Parler-only at the JS surface.
- */
-interface ParlerDescriptionFields {
-    description?: string;
-    voiceDescription?: string;
-    /** Parler voice-template field; also Supertonic's baked voice id. */
-    voice?: string;
-    pitch?: string;
-    expressivity?: string;
-    noise?: string;
-    reverb?: string;
-    quality?: string;
-}
-/**
- * Voice cloning reference. Audio8's codec encoder turns `referenceAudio` into
- * codes and they are prepended to the prompt as the speaker's own history, so
- * no speaker encoder and no enrolment step are involved. `referenceText` is
- * required alongside it there: the model conditions on it as the turn the
- * recording answers, and a wrong one degrades the clone. Accepted at
- * construction, on `reload()`, and per call (Audio8 only).
- */
-interface Audio8VoiceFields {
-    /**
-     * Chatterbox: voice-cloning reference audio path (wav). CosyVoice3:
-     * zero-shot / cross-lingual cloning reference (0.5-30 s hard limits,
-     * 5-15 s of clean speech recommended; multichannel input is downmixed to
-     * mono) — the native front-end tokenizes it (speech_tokenizer_v3),
-     * extracts the CAM++ speaker embedding and prompt mel at load, replacing
-     * the baked voice; requires the `cosyvoiceS3tokModel` +
-     * `cosyvoiceCampplusModel` files and fails the load (never silently falls
-     * back) when they are missing or the audio is unusable. Pair with
-     * `promptText` (the verbatim transcript) for zero-shot or omit it for
-     * cross-lingual. The reference is fixed at construction: `reload()`
-     * re-bakes the same recording (it is forwarded to the new addon instance)
-     * but cannot switch to a different one, so changing voices means a new
-     * instance. Audio8: the recording to clone, with `referenceText`
-     * alongside it. MOSS: the recording to clone, sampled at 24 kHz and
-     * fixed at construction; needs `files.mossCodecEncoder`.
-     */
-    referenceAudio?: string;
-    /** Audio8: what `referenceAudio` says. Required when cloning. */
-    referenceText?: string;
-}
-/**
- * MOSS-SoundEffect per-call generation controls (moss-sfx only). Each one is
- * optional; unset `steps` / `guidance` / `shift` use the defaults stored in the
- * model file (100 steps, guidance 4, shift 5).
- */
-interface MossSoundEffectFields {
-    /** Length of the sound effect, in (0, 30] seconds, rounded to 0.1 s. Defaults to 10. */
-    seconds?: number;
-    /** What the sound effect should avoid (classifier-free guidance negative). */
-    negativePrompt?: string;
-    /** Diffusion steps, 1..1000. More steps cost proportionally more time. */
-    steps?: number;
-    /** Classifier-free guidance scale, 1..50; 1 skips the negative branch. */
-    guidance?: number;
-    /** Flow-matching schedule shift, (0, 100]. */
-    shift?: number;
-}
-/** Mono PCM accepted by the MOSS-Speech engine: 16-bit integers or floats in [-1, 1]. */
-type MossSpeechPcm = Int16Array | Float32Array;
-/** One conversation turn for the MOSS-Speech engine: text or audio, never both. */
-interface MossSpeechMessage {
-    role: (typeof MOSS_SPEECH_ROLES)[number];
-    text?: string;
-    audio?: MossSpeechPcm;
-    /** Sample rate of `audio`, 8000..192000 Hz. */
-    sampleRate?: number;
-}
-/**
- * MOSS-Speech per-call fields (moss-speech only). The user turn is `audio`
- * (with `sampleRate`) or the run `input` text; `messages` carries the earlier
- * conversation.
- */
-interface MossSpeechFields {
-    /** The user's spoken turn, mono PCM. */
-    audio?: MossSpeechPcm;
-    /** Sample rate of `audio`, 8000..192000 Hz. */
-    sampleRate?: number;
-    /** Earlier turns, oldest first; the new user turn is appended after them. */
-    messages?: MossSpeechMessage[];
-    /** A system turn placed first; replaces the model's default system prompt. */
-    systemPrompt?: string;
-    /** Reply voice prompt (up to 60 s of mono PCM); the codec's default voice otherwise. */
-    replyVoice?: MossSpeechPcm;
-    /** Sample rate of `replyVoice`, 8000..192000 Hz. */
-    replyVoiceSampleRate?: number;
-    /** Reply with text only (no audio); the text comes back as `text`. */
-    textReply?: boolean;
-    /** Cut the spoken reply after this many seconds (0 = no cut). */
-    maxReplySeconds?: number;
-    /** Explicit bound on generated rows, 1..4096; omitted uses remaining model context. */
-    maxNewTokens?: number;
-    /** Greedy decoding instead of sampling. */
-    greedy?: boolean;
-    /** Sampling temperature, (0, 10] (default 0.7). */
-    temperature?: number;
-    /** Nucleus sampling, (0, 1] (default 0.95). */
-    topP?: number;
-    /** Top-k sampling, >= 0 (default 20). */
-    topK?: number;
-}
-/**
- * CosyVoice3 per-call instruction. Same values as the constructor's
- * `instruct`; a per-call conditioning control replaces the configured one for
- * that synthesis, and one control takes effect per synthesis.
- */
-interface CosyvoiceJobFields {
-    instruct?: string | CosyvoiceInstruct;
-}
-interface TTSGgmlOptions extends ParlerDescriptionFields, Audio8VoiceFields, TTSConditioningFields {
-    /** Pocket: generation/context controls; native sampling uses a portable RNG. */
-    maxTokens?: number;
-    noiseClamp?: number;
-    eosThreshold?: number;
-    framesAfterEos?: number;
-    files?: TTSGgmlFiles;
-    config?: TTSGgmlRuntimeConfig;
-    logger?: object;
-    lazySessionLoading?: boolean;
-    /** Explicit engine selection. Auto-detected from `files` when omitted. */
-    engine?: EngineType;
-    /** Chatterbox: directory of baked voice-conditioning tensors. */
-    voiceDir?: string;
-    /**
-     * RNG seed for Chatterbox CFM/SineGen, Supertonic latent generation,
-     * the MOSS-SoundEffect diffusion noise (0 when unset, so an unseeded prompt
-     * repeats its clip), or Pocket's portable sampling RNG. Pocket accepts
-     * integers from 0 to 4294967295 (inclusive).
-     */
-    seed?: number;
-    /**
-     * Move N layers to the GPU backend. Chatterbox: pass 99 to move everything.
-     * Supertonic: pass 99 to offload on GPU-capable hosts, including Android.
-     * Audio8: pass 99 to use Vulkan on Linux and Windows.
-     * MOSS: any non-zero value selects the GPU backend.
-     * CosyVoice3: pass 99 to offload on Metal (macOS / iOS), Vulkan (desktop
-     * Linux / Windows), or OpenCL/Adreno (Android); other hosts fall back to
-     * CPU by policy.
-     */
-    nGpuLayers?: number;
-    /**
-     * Chatterbox: cap on the T3 context length (prompt + generated speech
-     * tokens, 25 tokens ~= 1 second of audio). The KV cache is allocated up
-     * front at this length, so the cap directly bounds memory. Pass 0 to use
-     * the GGUF's full context; negative values are rejected.
-     * Pocket: FlowLM context capacity; accepts integers from 1 to 8192.
-     */
-    nCtx?: number;
-    /**
-     * Chatterbox-only T3 KV-cache storage dtype: `f32` | `f16` | `q8_0`.
-     * `f16` is the safe cross-backend default. `q8_0` is smaller and faster on
-     * supported backends, but is opt-in because not every backend implements
-     * its required operations.
-     */
-    kvCacheType?: "f32" | "f16" | "q8_0";
-    /** Override `std::thread::hardware_concurrency()`. */
-    threads?: number;
-    /**
-     * Native streaming chunk size; 0 disables. Chatterbox / CosyVoice3: speech
-     * tokens per chunk. MOSS: codec frames per chunk (12.5 per second).
-     * Supertonic: text tokens (Unicode code points) per chunk, about 50 for
-     * English (CJK is denser, 25-30). Supertonic native streaming cannot be
-     * combined with the LavaSR enhancer or denoiser.
-     */
-    streamChunkTokens?: number;
-    /**
-     * Chatterbox / CosyVoice3 / Supertonic smaller first chunk for low
-     * first-audio-out latency.
-     */
-    streamFirstChunkTokens?: number;
-    /**
-     * Supertonic-only: boundary-snap window for clause and whitespace chunk
-     * boundaries, as a percentage of the chunk target (engine default 20).
-     */
-    streamChunkTolerancePct?: number;
-    /**
-     * Supertonic-only: floor on every streamed chunk's size in text tokens
-     * (engine default 30); shorter trailing chunks merge into the previous one.
-     */
-    streamMinChunkTokens?: number;
-    /**
-     * Chatterbox: sliding-window S3Gen for native streaming — keep only this many
-     * already-emitted speech tokens of left context per chunk, which bounds the
-     * per-chunk cost on long utterances (typical 25-50; 0 = cumulative prefix).
-     * CosyVoice3: reserved / not yet effective — the pinned tts-cpp engine
-     * accepts the value but does not read it.
-     */
-    streamLeftContextTokens?: number;
-    /**
-     * Chatterbox CFM Euler step count for native streaming chunks. CosyVoice3:
-     * reserved / not yet effective — the engine runs a fixed 10-step schedule
-     * and ignores this.
-     */
-    cfmSteps?: number;
-    /**
-     * Chatterbox-only CFM Euler step count for batch synthesis; 0 = library
-     * default (2-step meanflow on Turbo, the model's full schedule on
-     * multilingual). The multilingual engine does not floor this, so values well
-     * below its schedule under-integrate.
-     */
-    batchCfmSteps?: number;
-    /**
-     * Chatterbox-only cap on speech tokens T3 decodes per synthesis (25 tokens ~=
-     * 1 second); engine default 1000 (~40 s). Longer single-call audio also needs
-     * `nCtx` to cover prompt plus generated tokens, or use `maxSentenceChars`.
-     */
-    nPredict?: number;
-    /**
-     * Chatterbox-only sentence auto-split: run T3 + S3Gen per segment of at most
-     * this many bytes (the reference CLI uses 180; 0 = off). Bounds each T3
-     * sequence, so `nPredict` applies per segment.
-     */
-    maxSentenceChars?: number;
-    /**
-     * Chatterbox-only raised-cosine crossfade between auto-split batch segments,
-     * in milliseconds (engine default 30). Native streaming stays gapless.
-     */
-    crossfadeMs?: number;
-    /** Chatterbox-only T3 repetition penalty (engine default 1.2; 1 = off). */
-    repeatPenalty?: number;
-    /**
-     * Chatterbox multilingual only: prosody intensity, above 0 more expressive,
-     * below 0 flatter (engine default 0.5). The Turbo variant ignores it.
-     */
-    exaggeration?: number;
-    /**
-     * Chatterbox multilingual only: T3 classifier-free guidance scale (engine
-     * default 0.5; 0 disables). Separate from `cfgRate`, which steers S3Gen.
-     */
-    cfgWeight?: number;
-    /** Chatterbox multilingual only: min-p sampling threshold (0 = off). */
-    minP?: number;
-    /**
-     * Chatterbox-only S3Gen classifier-free-guidance rate. The diffusion loop
-     * normally runs a batched conditioned + unconditioned pass combined by this
-     * rate. `0` skips the unconditioned pass; a positive value overrides the
-     * model's baked rate. Omit it to retain the baked rate.
-     */
-    cfgRate?: number;
-    /**
-     * CosyVoice3: verbatim transcript of `referenceAudio`, selecting the
-     * cloning mode per the upstream frontends — set it for zero-shot (the LM
-     * is prompted with transcript + reference speech tokens; best fidelity in
-     * the reference's own language), omit it for cross-lingual (timbre-only
-     * conditioning; best when synthesizing a different language than the
-     * reference). Without `referenceAudio` it still overrides the baked
-     * voice's transcript metadata for the LM prompt.
-     */
-    promptText?: string;
-    /**
-     * CosyVoice3: natural-language control (instruct2) — Chinese dialect, emotion,
-     * speed, volume, or style. Pass a structured object (e.g. `{ dialect:
-     * 'cantonese' }`, `{ emotion: 'happy' }`) which renders to the trained
-     * instruction, or a raw string for an arbitrary instruction. Applied on top of
-     * the selected voice's timbre; one control takes effect per synthesis.
-     */
-    instruct?: string | CosyvoiceInstruct;
-    /**
-     * Supertonic voice id baked into the GGUF, such as `F1` or `M1`. CosyVoice3:
-     * reserved / not yet effective — named-voice selection is not yet wired.
-     */
-    voice?: string;
-    /** Alias for `voice` for compatibility with `@qvac/tts-onnx`. */
-    voiceName?: string;
-    /**
-     * Supertonic-only: external voice JSON (`{ style_ttl, style_dp }`, e.g. a
-     * cloned voice) that overrides the baked `voice` preset. The engine checks
-     * the tensor sizes against the loaded model.
-     */
-    voiceJsonPath?: string;
-    /**
-     * Supertonic-only: text synthesized once at load so GPU pipelines (and a
-     * Core ML vocoder sidecar) are ready before the first `run()`; skipped on a
-     * plain CPU run. Use a sentence of representative length. Wins over the
-     * `vulkanCacheDir` default pre-warm.
-     */
-    prewarmText?: string;
-    /**
-     * Supertonic / CosyVoice3 Vulkan adapter index, also used by their LavaSR
-     * enhancer: 0 (default) = first adapter, N = the Nth, -1 = auto-pick by free
-     * VRAM preferring a discrete adapter. Only consulted on Vulkan.
-     */
-    vulkanDevice?: number;
-    /**
-     * CosyVoice3-only: treat the baked-voice prompt frames as attention-only
-     * conditioning in the flow, for a faster flow whose output deviates slightly
-     * from the reference (engine default off).
-     */
-    flowCutPrompt?: boolean;
-    /** Supertonic CFM steps (0 uses GGUF default); Pocket sampling steps (1–64, default 1). */
-    steps?: number;
-    /** Alias for `steps` for compatibility with `@qvac/tts-onnx`. */
-    numInferenceSteps?: number;
-    /**
-     * Speech-rate / duration multiplier (1.0 = unchanged, less than 1 slower,
-     * greater than 1 faster). Supertonic scales its native duration predictor.
-     * Chatterbox applies pitch-preserving WSOLA time-stretch, bounded to
-     * [0.25, 4.0].
-     */
-    speed?: number;
-    /** Supertonic optional `.npy` initial-noise tensor path. */
-    noiseNpyPath?: string;
-    /**
-     * LavaSR neural speech enhancement. Opt-in CPU/GGML bandwidth extension to
-     * 48 kHz, enabled by a GGUF path here or through `files.lavasrEnhancer`.
-     * Works for every engine, including the native chunk streaming of
-     * Chatterbox, Parler and CosyVoice3.
-     */
-    enhancer?: LavaSREnhancerOptions;
-    /**
-     * LavaSR neural speech denoiser (UL-UNAS). Opt-in preprocessing that runs
-     * before the enhancer and preserves the sample rate. Enabled by a GGUF path
-     * here or through `files.lavasrDenoiser`; rejected with native chunk
-     * streaming (batch synthesis only).
-     */
-    denoiser?: LavaSRDenoiserOptions;
-    /**
-     * Directory the addon scans for dynamically loaded ggml backends, as given.
-     * Unset scans the addon's own: the one its platform package or the app
-     * installed next to the native module.
-     */
-    backendsDir?: string;
-    /** Directory where ggml-opencl persists its compiled program binary. */
-    openclCacheDir?: string;
-    /**
-     * Supertonic + `useGPU: true` only: directory where the Vulkan backend
-     * persists its compiled pipeline cache (`GGML_VK_PIPELINE_CACHE_DIR`).
-     * Unset means no cross-process cache or load-time pre-warm.
-     */
-    vulkanCacheDir?: string;
-    /** Chatterbox MTL MeCab/IPAdic dictionary directory for Japanese. */
-    mecabDictPath?: string;
-    mecabDictDir?: string;
-    /** Chatterbox MTL Cangjie TSV path for Chinese. */
-    cangjieTsvPath?: string;
-    /**
-     * Parler, Audio8 and Chatterbox sampling knobs; each unset defers to the
-     * engine's own defaults (Parler: temperature 1.0, top-k 50; Audio8:
-     * temperature 0.7, top-k 50, top-p 0.9; Chatterbox: temperature 0.8,
-     * top-k 1000, top-p 0.95, where temperature 0 decodes greedily and top-k 0
-     * disables the cutoff). Audio8 filters by top-k/top-p on the raw logits and
-     * only then applies the temperature, following its reference.
-     * Pocket: sampling temperature; accepts finite values from 0 to 10.
-     */
-    temperature?: number;
-    topK?: number;
-    topP?: number;
-    /**
-     * Generation-length cap in decoder frames; 0 = engine default. Parler runs
-     * ~86 frames/s, Audio8 ~21.5.
-     */
-    maxFrames?: number;
-    /** Audio8: take the argmax instead of sampling. */
-    greedy?: boolean;
-    /**
-     * MOSS: target length in codec frames (12.5 per second), from 0 to 2015
-     * (about 161 s); 0 or unset keeps the length free. Set at construction or
-     * with `reload()`, not per call.
-     */
-    durationTokens?: number;
-    /**
-     * MOSS-TTSD dialogue: one 24 kHz reference recording per speaker, in the
-     * order the text tags them (`[S1]`, `[S2]`, ...). The model continues the
-     * references, so the input text must open with each reference's transcript
-     * under its tag, followed by the lines to generate; sentence streaming is
-     * therefore rejected (use `run()` or `streamChunkTokens`). Needs
-     * `files.mossCodecEncoder`, excludes `referenceAudio`, and is fixed for the
-     * instance. With a `modelDir`, requires a `moss-ttsd-*.gguf` backbone.
-     */
-    dialogueReferences?: string[];
-    minNewTokens?: number;
-    /** Parler prompt digit expansion (engine default: enabled). */
-    normalizeNumbers?: boolean;
-    opts?: object;
-    exclusiveRun?: boolean;
-}
-interface InferenceState {
-    configLoaded: boolean;
-    weightsLoaded: boolean;
-    destroyed: boolean;
-}
-interface TTSOutputChunk {
-    chunkIndex?: number;
-    isLast?: boolean;
-    /** Signed 16-bit mono PCM audio payload. */
-    outputArray: Int16Array;
-    /** MOSS-Speech: the reply's text channel (the whole answer with `textReply`). */
-    text?: string;
-    /**
-     * Output sample rate. The native engine rate (24000 for Chatterbox,
-     * CosyVoice3 and MOSS; 44100 for Supertonic, Parler, and Audio8), or 48000
-     * when the LavaSR enhancer is active.
-     */
-    sampleRate?: number;
-}
-interface RuntimeStats {
-    firstAudioMs?: number;
-    totalTime: number;
-    tokensPerSecond: number;
-    realTimeFactor: number;
-    audioDurationMs: number;
-    totalSamples: number;
-    /** Active compute device after load-time backend selection. 0 = CPU, 1 = GPU. */
-    backendDevice?: number;
-    /** Stable backend code: 0=CPU, 1=Metal, 2=CUDA, 3=Vulkan, 4=OpenCL, 99=other GPU. */
-    backendId?: number;
-    /** LavaSR enhancer compute device. -1 = not loaded, 0 = CPU, 1 = GPU. */
-    enhancerBackendDevice?: number;
-    /** LavaSR enhancer backend code, using the same values as `backendId`. */
-    enhancerBackendId?: number;
-    /** LavaSR denoiser compute device. -1 = not loaded, 0 = CPU, 1 = GPU. */
-    denoiserBackendDevice?: number;
-    /** LavaSR denoiser backend code, using the same values as `backendId`. */
-    denoiserBackendId?: number;
-    /** 1 when a present GPU is unsupported by engine policy; 0 otherwise. */
-    gpuUnsupported?: number;
-    /**
-     * Audio8 and MOSS only: codec frames generated (Audio8 on a fixed 46 ms
-     * grid, MOSS at 12.5 per second). This is the unit their `tokensPerSecond`
-     * counts, in batch and in streaming alike.
-     */
-    generatedFrames?: number;
-    /**
-     * Audio8 only, macOS / iOS: 1 while an Apple Core ML sidecar for the codec's
-     * synthesis stack is attached (a compiled `audio8-codec-decoder.mlmodelc`
-     * next to the decoder GGUF); 0 without one, or once a failing sidecar has
-     * been retired. Streams report the last chunk that supplied this field.
-     */
-    codecSidecarLoaded?: number;
-    /**
-     * Audio8 only: 1 when this synthesis ran the codec's synthesis stack on the
-     * Apple Core ML sidecar -- a compiled `audio8-codec-decoder.mlmodelc` next
-     * to the decoder GGUF on macOS / iOS -- 0 when it ran on the ggml backend
-     * `backendId` reports (which the language model always uses). A loaded
-     * sidecar that cannot serve a call falls back to ggml and reports 0 for it.
-     * Streams report the last chunk that supplied this field, not a sum.
-     */
-    codecOnCoreml?: number;
-    /** Chatterbox only: T3 decode wall time of the last synthesis, in ms. */
-    t3Ms?: number;
-    /** Chatterbox only: S3Gen + HiFT wall time of the last synthesis, in ms. */
-    s3genMs?: number;
-    /** Chatterbox only: speech tokens T3 emitted in the last synthesis. */
-    t3Tokens?: number;
-    /**
-     * CosyVoice3 / Audio8: the engine's own wall time for the last synthesis, in
-     * ms (the sum of its stages), excluding the addon's post-processing.
-     */
-    stageTotalMs?: number;
-    /** CosyVoice3 stage wall times of the last synthesis, in ms. */
-    lmPrefillMs?: number;
-    lmDecodeMs?: number;
-    flowFrontendMs?: number;
-    ditEulerMs?: number;
-    hiftF0Ms?: number;
-    hiftSourceMs?: number;
-    hiftStftMs?: number;
-    hiftDecodeMs?: number;
-    /**
-     * CosyVoice3 work counters of the last synthesis: LM decode steps, speech
-     * tokens generated, flow frames in (prompt + generated), mel frames out,
-     * text ids in the LM prefill, and prompt speech tokens the LM continued from.
-     */
-    decodeSteps?: number;
-    speechTokens?: number;
-    flowFrames?: number;
-    melFrames?: number;
-    textIds?: number;
-    promptSpeechTokens?: number;
-    /** Audio8 stage wall times of the last synthesis, in ms. */
-    voiceEncodeMs?: number;
-    promptMs?: number;
-    prefillMs?: number;
-    sampleMs?: number;
-    fastDecodeMs?: number;
-    slowDecodeMs?: number;
-    codecLatentMs?: number;
-    codecSynthMs?: number;
-    resampleMs?: number;
-}
-/**
- * tts-cpp's canonical conditioning vocabulary and each engine's supported
- * subset, keyed by engine type (`TTSGgml.getVoiceControls()`).
- */
-interface VoiceControlsCatalog {
-    emotions: string[];
-    paces: string[];
-    engines: Record<string, {
-        emotions: string[];
-        paces: string[];
-    }>;
-}
-interface SentenceStreamChunkMeta {
-    chunkIndex?: number;
-    sentenceChunk?: string;
-    /**
-     * True on the final pre-chunked synthesis output. Undefined for async
-     * iterator streaming where the final chunk is not known up front.
-     */
-    isLast?: boolean;
-}
-interface SentenceStreamOptions extends ParlerDescriptionFields, Audio8VoiceFields, TTSConditioningFields, CosyvoiceJobFields {
-    /** BCP-47 locale for `Intl.Segmenter` when available. */
-    locale?: string;
-    /** Maximum graphemes per chunk; defaults to 300, or 120 for Korean. */
-    maxChunkScalars?: number;
-}
-interface RunStreamingOptions extends ParlerDescriptionFields, Audio8VoiceFields, TTSConditioningFields, CosyvoiceJobFields {
-    accumulateSentences?: boolean;
-    sentenceDelimiter?: RegExp;
-    sentenceDelimiterPreset?: SentenceDelimiterPreset;
-    maxBufferScalars?: number;
-    flushAfterMs?: number;
-}
-/** Input accepted by `runStreaming`. */
-type TextStreamInput = string | string[] | Iterable<string> | AsyncIterable<string>;
-interface TTSRunInput extends ParlerDescriptionFields, Audio8VoiceFields, TTSConditioningFields, CosyvoiceJobFields, MossSoundEffectFields, MossSpeechFields {
-    type?: string;
-    input: string;
-    streamOutput?: boolean;
-    locale?: string;
-    maxChunkScalars?: number;
-    /**
-     * Cancels non-streaming `run()`. An already-aborted signal rejects without
-     * native dispatch. Ignored by all streaming paths.
-     */
-    signal?: AbortSignal;
-}
+import type * as Fit from "./lib/fit";
+import { type EngineType } from "./lib/engineIds";
+import type * as PublicTypes from "./lib/types";
+import type { InferenceState, RunStreamingOptions, SentenceStreamChunkMeta, SentenceStreamOptions, TextStreamInput, TTSGgmlOptions, TTSOutputChunk, TTSRunInput, VoiceControlsCatalog } from "./lib/types";
 /**
  * GGML-backed TTS via the `tts-cpp` library. Wraps the chatterbox,
  * supertonic, parler, cosyvoice3, audio8 and moss engines behind a single
@@ -848,190 +34,15 @@ declare class TTSGgml {
     addon: unknown;
     private readonly _job;
     private readonly _runExclusive;
-    private _ttsInferenceQueueWaiter;
-    private _sentenceStreamCtx;
-    private _config;
-    private _pocketJobPending;
-    private _pocketCancelPromise;
-    private _pocketLifecycleInProgress;
-    private _pocketParams;
-    private _pocketOptions;
-    private _pocketFiles;
+    private readonly _exclusiveStreams;
+    /** Everything resolved from the constructor options; reload rewrites parts of it. */
+    private readonly _setup;
+    private readonly _sentenceStream;
+    private readonly _events;
+    /** The Pocket execution rules, or null for every other engine. */
+    private readonly _pocket;
     private _lazySessionLoading;
-    private _outputSampleRate;
-    private _engineType;
-    private _voicesDir?;
-    private _supertonicModelPath?;
-    private _t3ModelPath?;
-    private _s3genModelPath?;
-    private _cosyvoiceModelDir?;
-    private _cosyvoiceLlmModelPath?;
-    private _cosyvoiceFlowModelPath?;
-    private _cosyvoiceHiftModelPath?;
-    private _cosyvoiceS3tokModelPath?;
-    private _cosyvoiceCampplusModelPath?;
-    private _mecabDictPath?;
-    private _cangjieTsvPath?;
-    private _referenceAudio?;
-    private _voiceDir?;
-    private _seed?;
-    private _nGpuLayers?;
-    private _nCtx?;
-    private _kvCacheType?;
-    private _threads?;
-    private _streamChunkTokens?;
-    private _streamFirstChunkTokens?;
-    private _streamLeftContextTokens?;
-    private _cfmSteps?;
-    private _cfgRate?;
-    private _batchCfmSteps?;
-    private _nPredict?;
-    private _maxSentenceChars?;
-    private _crossfadeMs?;
-    private _repeatPenalty?;
-    private _exaggeration?;
-    private _cfgWeight?;
-    private _minP?;
-    private _streamChunkTolerancePct?;
-    private _streamMinChunkTokens?;
-    private _voiceJsonPath?;
-    private _prewarmText?;
-    private _vulkanDevice?;
-    private _flowCutPrompt?;
-    private _cosyvoiceVocabPath?;
-    private _cosyvoiceMergesPath?;
-    private _cosyvoiceVoiceModelPath?;
-    private _promptText?;
-    private _instruct?;
-    private _voice?;
-    private _steps?;
-    private _speed?;
-    private _noiseNpyPath?;
-    private _enhancerGgufPath?;
-    private _denoiserGgufPath?;
-    private _backendsDir?;
-    private _openclCacheDir?;
-    private _vulkanCacheDir?;
-    private _parlerModelPath?;
-    private _audio8LmPath?;
-    private _audio8CodecDecoderPath?;
-    private _audio8CodecEncoderPath?;
-    private _mossBackbonePath?;
-    private _mossCodecDecoderPath?;
-    private _mossCodecEncoderPath?;
-    private _mossSoundEffectPath?;
-    private _mossSpeechModelPath?;
-    private _mossSpeechCodecPath?;
-    private _dialogueReferences?;
-    private _durationTokens?;
-    private _referenceText?;
-    private _greedy?;
-    private _description?;
-    private _emotion?;
-    private _pitch?;
-    private _pace?;
-    private _expressivity?;
-    private _noise?;
-    private _reverb?;
-    private _quality?;
-    private _temperature?;
-    private _topK?;
-    private _topP?;
-    private _maxFrames?;
-    private _minNewTokens?;
-    private _normalizeNumbers?;
     constructor(options?: TTSGgmlOptions);
-    private _resolveEngineAndModelPaths;
-    private _resolveAudio8ModelPaths;
-    private _mossBackbonePatterns;
-    private _findMossBackbone;
-    private _resolveMossModelPaths;
-    private _assignMossVoiceOptions;
-    private _assignSynthesisOptions;
-    private _assertEngineStreamingSupport;
-    private _requestsChunkStreaming;
-    private _assertParlerOptionConsistency;
-    private _assertSamplerOptionSupport;
-    /**
-     * Options only some engines read. Set on any other engine they would be
-     * dropped without effect, so they are rejected by name instead.
-     */
-    private _assertEngineScopedOptions;
-    private _assertAudio8OptionConsistency;
-    private _assertMossSoundEffectOptionConsistency;
-    private _assertMossSpeechOptionConsistency;
-    private _assertMossSpeechOutputRate;
-    private _assertMossSoundEffectOutputRate;
-    private _assertMossOptionConsistency;
-    private _assertNoMossOnlyOptions;
-    private _assertMossDialogueReferences;
-    private _assertSentenceStreamingAllowed;
-    private _assertMossOutputRate;
-    private _assertMossVoiceConsistent;
-    /**
-     * A recording without its transcript is accepted by the model but degrades
-     * the clone silently, and a transcript alone has nothing to attach to, so
-     * both halves have to arrive together.
-     */
-    private _assertAudio8VoiceConsistent;
-    /**
-     * `promptText` is deliberately not required: its absence selects
-     * cross-lingual mode. With a model dir present but no cloning GGUFs in it,
-     * the native load fail-closes instead.
-     */
-    private _assertCosyvoiceCloneConsistent;
-    private _assertCosyvoiceOptionConsistency;
-    /**
-     * Validate the cross-engine emotion/pace surface against this engine, and
-     * enforce CosyVoice3's one-instruction-per-synthesis rule.
-     */
-    private _assertConditioningConsistency;
-    /**
-     * The per-call surface of a non-Parler engine: only the cross-engine
-     * emotion/pace it declares support for, never Parler's template fields.
-     */
-    private _resolveConditioningJobFields;
-    /**
-     * CosyVoice3's per-call instruction, rendered like the constructor's. The
-     * native layer reads it next to emotion/pace, and any of the three replaces
-     * the configured conditioning for that synthesis.
-     */
-    private _resolveJobInstruct;
-    /**
-     * Parler's per-call surface: description/template fields, where a per-call
-     * template cannot be merged with a constructor-level free-text description.
-     */
-    private _resolveParlerJobFields;
-    /**
-     * The voice a per-call override actually synthesizes with. Mirrors
-     * Audio8Model::resolveVoice: a per-call recording replaces both halves, so
-     * it cannot inherit the configured transcript, which describes a different
-     * recording; a per-call transcript alone corrects the configured one.
-     */
-    private _mergeAudio8Voice;
-    /**
-     * Extract + validate the per-call Audio8 voice fields from a run input or
-     * streaming options. Returns undefined when none are present.
-     */
-    private _resolveAudio8JobFields;
-    private _assertNoSpeechFields;
-    private _assertNoOtherEngineFields;
-    /**
-     * Validate the MOSS-Speech call and fold the new user turn (spoken `audio`
-     * or text `input`) and `systemPrompt` into the native `messages` list.
-     */
-    private _resolveSpeechJobFields;
-    /**
-     * Extract + validate the per-call MOSS-SoundEffect controls from a run
-     * input. Returns undefined when none are present.
-     */
-    private _resolveSoundEffectJobFields;
-    /**
-     * The per-call fields of whichever engine is loaded, if any are set. Parler
-     * takes the full description/template surface, Audio8 its voice override,
-     * and every engine the cross-engine conditioning it supports.
-     */
-    private _resolveJobFields;
     getEngineType(): EngineType;
     getApiDefinition(): string;
     getState(): InferenceState;
@@ -1058,95 +69,23 @@ declare class TTSGgml {
      * small streamed fragments are coalesced.
      */
     runStreaming(textStream: TextStreamInput, options?: RunStreamingOptions): Promise<QvacResponse<TTSOutputChunk & SentenceStreamChunkMeta>>;
-    private _enqueueExclusiveTtsResponse;
-    private _resolveRunStreamingOptions;
-    private _normalizeTextStream;
-    private _runTextStreamOrchestrator;
-    private _sentenceStreamTextIterableDrive;
-    /**
-     * Audio8 and MOSS report `tokensPerSecond` as codec frames per second, so
-     * the streaming aggregate has to count frames too rather than characters.
-     */
-    private _pacesOnFrames;
+    /** Streams line up behind each other under `exclusiveRun`; otherwise they start at once. */
+    private _startStream;
     private _runStreamOrchestrator;
-    private _sentenceStreamDriveBody;
     private _load;
     private _buildTtsParams;
-    private _buildCosyvoiceParams;
-    private _buildChatterboxParams;
-    private _buildSupertonicParams;
-    private _buildParlerParams;
-    private _buildAudio8Params;
-    private _buildMossParams;
-    private _buildMossSoundEffectParams;
-    private _buildMossSpeechParams;
-    /** The knobs every engine in SAMPLING_ENGINES reads. */
-    private _assignSamplingParams;
-    /**
-     * Backend and output plumbing for the engines that take no `language`,
-     * where `_assignCommonNativeParams` would be the wrong shape.
-     */
-    private _assignBackendParams;
-    private _assignCommonNativeParams;
-    /** LavaSR post-processing paths, shared by every engine that supports them. */
-    private _assignLavasrParams;
+    /** A native instance that reports through this wrapper's output callback. */
+    private _spawnAddon;
     private _createAddon;
     unload(): Promise<void>;
     private _unloadModel;
     destroy(): Promise<void>;
+    private _withPocketLifecycle;
     private _runInternal;
-    private _mergeSentenceStreamStats;
-    private _rejectActiveChunk;
-    private _endJobWithStats;
     private _addonOutputCallback;
-    private _handleAddonError;
-    private _handleAddonOutput;
-    private _enrichStreamChunk;
-    private _handleAddonStats;
-    private _waitPocketCancel;
-    private _withPocketLifecycle;
-    private _checkPocketReload;
-    private _checkPocketRequest;
     private _dispatchJob;
     cancel(): Promise<void>;
-    private _failAndClearActiveResponse;
-    /** Everything reload() may overwrite, so a rejected reload can undo itself. */
-    private _captureReloadableState;
-    private _restoreReloadableState;
-    private _applyReloadableRuntimeConfig;
-    private _assertMossReloadKeepsVoice;
-    private _applyMossReload;
-    private _applyReloadableConditioning;
-    private _applyReloadableParlerConfig;
-    /**
-     * Apply the new configuration and build the native parameters from it. A
-     * rejected value leaves the instance exactly as it was, so a later partial
-     * reload is not validated against state the caller never accepted.
-     */
-    private _applyReloadableConfig;
-    private _reloadPocket;
     reload(newConfig?: Record<string, unknown>): Promise<void>;
-    /**
-     * The voice a reload lands on. Same rule as _mergeAudio8Voice and
-     * Audio8Model::resolveVoice: a new recording replaces both halves, because
-     * the configured transcript describes the recording being replaced. Reload
-     * reads `undefined` as "not supplied", so an explicit empty string reaches
-     * the guard instead of being ignored.
-     */
-    private _mergeAudio8ReloadVoice;
-    /**
-     * The knobs a reload lands on, merged the same way as the voice so the
-     * whole set can be checked before any of it is written.
-     */
-    private _mergeAudio8ReloadSampling;
-    private _applyAudio8Sampling;
-    /**
-     * Audio8 voice + sampling knobs are reloadable; they rebuild the engine's
-     * sampler and speaker history. Both merges are checked before either is
-     * written, so a rejected reload leaves the instance exactly as it was
-     * rather than half-moved onto the configuration that was refused.
-     */
-    private _applyAudio8Reload;
     static getModelKey(_params?: unknown): string;
     /**
      * tts-cpp's canonical emotion / pace vocabulary and each engine's supported
@@ -1159,61 +98,35 @@ declare class TTSGgml {
     private _optionalAddon;
     private _getLogger;
 }
-type NamespaceFiles = TTSGgmlFiles;
-type NamespaceRuntimeConfig = TTSGgmlRuntimeConfig;
-type NamespaceOptions = TTSGgmlOptions;
-type NamespaceEnhancerOptions = LavaSREnhancerOptions;
-type NamespaceDenoiserOptions = LavaSRDenoiserOptions;
-type NamespaceRuntimeStats = RuntimeStats;
-type NamespaceOutputChunk = TTSOutputChunk;
-type NamespaceSentenceStreamChunkMeta = SentenceStreamChunkMeta;
-type NamespaceSentenceStreamOptions = SentenceStreamOptions;
-type NamespaceRunStreamingOptions = RunStreamingOptions;
-type NamespaceTextStreamInput = TextStreamInput;
-type NamespaceRunInput = TTSRunInput;
-type NamespaceInferenceState = InferenceState;
-type NamespaceCosyvoiceInstruct = CosyvoiceInstruct;
-type NamespaceVoiceControlsCatalog = VoiceControlsCatalog;
-type NamespaceFitEngine = TtsFitEngine;
-type NamespaceFitRequest = TtsFitRequest;
-type NamespaceFitResult = TtsFitResult;
-type NamespaceFitStatus = TtsFitStatus;
-type NamespaceSupertonicFit = SupertonicFitRequest;
-type NamespaceParlerFit = ParlerFitRequest;
-type NamespaceChatterboxFit = ChatterboxFitRequest;
-type NamespaceAudio8Fit = Audio8FitRequest;
-type NamespaceMossFit = MossFitRequest;
-type NamespaceMossSoundEffectFit = MossSoundEffectFitRequest;
-type NamespaceCosyvoiceFit = CosyvoiceFitRequest;
 declare namespace TTSGgml {
     export import QvacErrorAddonTTSGgml = errorModule.QvacErrorAddonTTSGgml;
     export import ERR_CODES = errorModule.ERR_CODES;
-    type TTSGgmlFiles = NamespaceFiles;
-    type TTSGgmlRuntimeConfig = NamespaceRuntimeConfig;
-    type TTSGgmlOptions = NamespaceOptions;
-    type LavaSREnhancerOptions = NamespaceEnhancerOptions;
-    type LavaSRDenoiserOptions = NamespaceDenoiserOptions;
-    type RuntimeStats = NamespaceRuntimeStats;
-    type TTSOutputChunk = NamespaceOutputChunk;
-    type SentenceStreamChunkMeta = NamespaceSentenceStreamChunkMeta;
-    type SentenceStreamOptions = NamespaceSentenceStreamOptions;
-    type RunStreamingOptions = NamespaceRunStreamingOptions;
-    type TextStreamInput = NamespaceTextStreamInput;
-    type TTSRunInput = NamespaceRunInput;
-    type InferenceState = NamespaceInferenceState;
-    type CosyvoiceInstruct = NamespaceCosyvoiceInstruct;
-    type VoiceControlsCatalog = NamespaceVoiceControlsCatalog;
-    type TtsFitEngine = NamespaceFitEngine;
-    type TtsFitRequest = NamespaceFitRequest;
-    type TtsFitResult = NamespaceFitResult;
-    type TtsFitStatus = NamespaceFitStatus;
-    type SupertonicFitRequest = NamespaceSupertonicFit;
-    type ParlerFitRequest = NamespaceParlerFit;
-    type ChatterboxFitRequest = NamespaceChatterboxFit;
-    type Audio8FitRequest = NamespaceAudio8Fit;
-    type MossFitRequest = NamespaceMossFit;
-    type MossSoundEffectFitRequest = NamespaceMossSoundEffectFit;
-    type CosyvoiceFitRequest = NamespaceCosyvoiceFit;
-    const assessFit: typeof assessFitImpl;
+    type TTSGgmlFiles = PublicTypes.TTSGgmlFiles;
+    type TTSGgmlRuntimeConfig = PublicTypes.TTSGgmlRuntimeConfig;
+    type TTSGgmlOptions = PublicTypes.TTSGgmlOptions;
+    type LavaSREnhancerOptions = PublicTypes.LavaSREnhancerOptions;
+    type LavaSRDenoiserOptions = PublicTypes.LavaSRDenoiserOptions;
+    type RuntimeStats = PublicTypes.RuntimeStats;
+    type TTSOutputChunk = PublicTypes.TTSOutputChunk;
+    type SentenceStreamChunkMeta = PublicTypes.SentenceStreamChunkMeta;
+    type SentenceStreamOptions = PublicTypes.SentenceStreamOptions;
+    type RunStreamingOptions = PublicTypes.RunStreamingOptions;
+    type TextStreamInput = PublicTypes.TextStreamInput;
+    type TTSRunInput = PublicTypes.TTSRunInput;
+    type InferenceState = PublicTypes.InferenceState;
+    type CosyvoiceInstruct = PublicTypes.CosyvoiceInstruct;
+    type VoiceControlsCatalog = PublicTypes.VoiceControlsCatalog;
+    type TtsFitEngine = Fit.TtsFitEngine;
+    type TtsFitRequest = Fit.TtsFitRequest;
+    type TtsFitResult = Fit.TtsFitResult;
+    type TtsFitStatus = Fit.TtsFitStatus;
+    type SupertonicFitRequest = Fit.SupertonicFitRequest;
+    type ParlerFitRequest = Fit.ParlerFitRequest;
+    type ChatterboxFitRequest = Fit.ChatterboxFitRequest;
+    type Audio8FitRequest = Fit.Audio8FitRequest;
+    type MossFitRequest = Fit.MossFitRequest;
+    type MossSoundEffectFitRequest = Fit.MossSoundEffectFitRequest;
+    type CosyvoiceFitRequest = Fit.CosyvoiceFitRequest;
+    const assessFit: typeof Fit.assessFit;
 }
 export = TTSGgml;