@qvac/sdk 0.15.0 → 0.17.0

package/dist/schemas/load-model.js

@@ -4,13 +4,14 @@ import { whisperConfigSchema, parakeetLoadConfigSchema } from './transcription-c
 import { bciConfigSchema } from './bci-config.js';
 import { delegateSchema } from './delegate.js';
 import { nmtConfigBaseSchema, nmtConfigSchema } from './translation-config.js';
-import { LEGACY_TTS_ONNX_MODEL_CONFIG_FIELDS, ttsChatterboxLoadConfigSchema, ttsConfigSchema, ttsSupertonicLoadConfigSchema } from './text-to-speech.js';
+import { LEGACY_TTS_ONNX_MODEL_CONFIG_FIELDS, ttsChatterboxLoadConfigSchema, ttsConfigSchema, ttsParlerLoadConfigSchema, ttsSupertonicLoadConfigSchema } from './text-to-speech.js';
 import { ocrConfigSchema } from './ocr.js';
 import { modelSrcInputSchema, modelInputToSrcSchema, modelInputToNameSchema } from './model-src-utils.js';
-import { llmModelTypeSchema, whisperModelTypeSchema, bciModelTypeSchema, parakeetModelTypeSchema, embeddingsModelTypeSchema, nmtModelTypeSchema, ttsModelTypeSchema, ocrModelTypeSchema, diffusionModelTypeSchema, vlaModelTypeSchema, classificationModelTypeSchema, ModelType, ModelTypeAliases, normalizeModelType } from './model-types.js';
+import { llmModelTypeSchema, whisperModelTypeSchema, bciModelTypeSchema, parakeetModelTypeSchema, embeddingsModelTypeSchema, nmtModelTypeSchema, ttsModelTypeSchema, ocrModelTypeSchema, diffusionModelTypeSchema, audioGenModelTypeSchema, vlaModelTypeSchema, classificationModelTypeSchema, ModelType, ModelTypeAliases, normalizeModelType } from './model-types.js';
 import { sdcppConfigSchema } from './sdcpp-config.js';
 import { vlaConfigSchema } from './vla.js';
 import { classificationConfigSchema } from './classification.js';
+import { audioGenConfigSchema } from '../schemas/audio-gen.js';
 // Set of all built-in model types (canonical + aliases) for catch-all exclusion
 const builtInModelTypes = new Set([...Object.values(ModelType), ...Object.keys(ModelTypeAliases)]);
 export function isBuiltInModelType(modelType) {
@@ -43,10 +44,11 @@ const modelConfigKeysByModelType = new Map([
     [ModelType.nmtcppTranslation, configKeys(...nmtConfigBaseSchema.options)],
     [
         ModelType.ttsGgml,
-        configKeys(ttsChatterboxLoadConfigSchema, ttsSupertonicLoadConfigSchema, LEGACY_TTS_ONNX_MODEL_CONFIG_FIELDS)
+        configKeys(ttsChatterboxLoadConfigSchema, ttsSupertonicLoadConfigSchema, ttsParlerLoadConfigSchema, LEGACY_TTS_ONNX_MODEL_CONFIG_FIELDS)
     ],
     [ModelType.ggmlOcr, configKeys(ocrConfigSchema)],
     [ModelType.sdcppGeneration, configKeys(sdcppConfigSchema)],
+    [ModelType.audiogenGgml, configKeys(audioGenConfigSchema)],
     [ModelType.ggmlVla, configKeys(vlaConfigSchema)],
     [ModelType.ggmlClassification, configKeys(classificationConfigSchema)]
 ]);
@@ -142,6 +144,14 @@ export const loadBuiltinModelOptionsBaseSchema = z.union([
         modelConfig: sdcppConfigSchema.strict().optional()
     })
         .strict(),
+    z
+        .object({
+        ...loadModelCommonFields,
+        modelSrc: modelSrcInputSchema.optional(),
+        modelType: audioGenModelTypeSchema,
+        modelConfig: audioGenConfigSchema
+    })
+        .strict(),
     z
         .object({
         ...loadModelCommonFields,
@@ -345,6 +355,25 @@ export const loadBuiltinToRequestSchema = z.discriminatedUnion('modelType', [
         delegate: data.delegate,
         ...(data.requestId !== undefined && { requestId: data.requestId })
     })),
+    z
+        .object({
+        ...loadModelRequestCommonFields,
+        modelSrc: modelSrcInputSchema.optional(),
+        modelType: audioGenModelTypeSchema,
+        modelConfig: audioGenConfigSchema
+    })
+        .strict()
+        .transform((data) => ({
+        type: 'loadModel',
+        modelType: ModelType.audiogenGgml,
+        modelSrc: data.modelSrc ? modelInputToSrcSchema.parse(data.modelSrc) : '',
+        modelName: data.modelSrc ? modelInputToNameSchema.parse(data.modelSrc) : undefined,
+        modelConfig: data.modelConfig,
+        seed: data.seed ?? false,
+        withProgress: data.withProgress ?? !!data.onProgress,
+        delegate: data.delegate,
+        ...(data.requestId !== undefined && { requestId: data.requestId })
+    })),
     z
         .object({
         ...loadModelRequestCommonFields,
@@ -477,6 +506,12 @@ export const loadDiffusionModelRequestSchema = commonModelConfigSchema
     modelConfig: sdcppConfigSchema.optional()
 })
     .strict();
+export const loadAudioGenModelRequestSchema = commonModelConfigSchema
+    .extend({
+    modelType: z.literal(ModelType.audiogenGgml),
+    modelConfig: audioGenConfigSchema
+})
+    .strict();
 export const loadVlaModelRequestSchema = commonModelConfigSchema
     .extend({
     modelType: z.literal(ModelType.ggmlVla),
@@ -490,12 +525,14 @@ export const loadClassificationModelRequestSchema = commonModelConfigSchema
 })
     .strict();
 // Custom plugin catch-all: accepts any modelType string EXCEPT built-ins
-export const loadCustomPluginModelRequestSchema = commonModelConfigSchema.extend({
+export const loadCustomPluginModelRequestSchema = commonModelConfigSchema
+    .extend({
     modelType: z.string().refine((val) => !builtInModelTypes.has(val), {
         message: 'Built-in model types must use their specific schema'
     }),
     modelConfig: z.record(z.string(), z.unknown()).optional()
-});
+})
+    .meta({ title: 'LoadModelCustomPluginRequest' });
 // Union of all load model request types (using z.union since each modelType accepts multiple values)
 export const loadModelSrcRequestSchema = z
     .union([
@@ -508,6 +545,7 @@ export const loadModelSrcRequestSchema = z
     loadTtsModelRequestSchema,
     loadOcrModelRequestSchema,
     loadDiffusionModelRequestSchema,
+    loadAudioGenModelRequestSchema,
     loadVlaModelRequestSchema,
     loadClassificationModelRequestSchema,
     loadCustomPluginModelRequestSchema
@@ -515,7 +553,8 @@ export const loadModelSrcRequestSchema = z
     .transform((data) => ({
     ...data,
     seed: data.seed ?? false
-}));
+}))
+    .meta({ title: 'LoadModelSrcRequest' });
 // Combined request schema: load new model OR reload config
 export const loadModelRequestSchema = z.union([
     loadModelSrcRequestSchema,

package/dist/schemas/mcp-adapter.d.ts

@@ -1,4 +1,4 @@
-import type { JsonSchema } from './json-schema';
+import type { JsonSchema } from '../schemas/json-schema.js';
 export type McpTool = {
     name: string;
     description?: string | undefined;

package/dist/schemas/model-src-utils.d.ts

@@ -13,6 +13,7 @@ export declare const modelDescriptorSchema: z.ZodObject<{
     addon: z.ZodOptional<z.ZodUnion<readonly [z.ZodEnum<{
         embeddings: "embeddings";
         diffusion: "diffusion";
+        audiogen: "audiogen";
         tts: "tts";
         ocr: "ocr";
         vla: "vla";
@@ -26,6 +27,7 @@ export declare const modelDescriptorSchema: z.ZodObject<{
         "parakeet-transcription": "parakeet-transcription";
         "ggml-ocr": "ggml-ocr";
         "sdcpp-generation": "sdcpp-generation";
+        "audiogen-ggml": "audiogen-ggml";
         "ggml-vla": "ggml-vla";
         "ggml-classification": "ggml-classification";
         llm: "llm";
@@ -50,6 +52,7 @@ export declare const modelSrcInputSchema: z.ZodUnion<readonly [z.ZodString, z.Zo
     addon: z.ZodOptional<z.ZodUnion<readonly [z.ZodEnum<{
         embeddings: "embeddings";
         diffusion: "diffusion";
+        audiogen: "audiogen";
         tts: "tts";
         ocr: "ocr";
         vla: "vla";
@@ -63,6 +66,7 @@ export declare const modelSrcInputSchema: z.ZodUnion<readonly [z.ZodString, z.Zo
         "parakeet-transcription": "parakeet-transcription";
         "ggml-ocr": "ggml-ocr";
         "sdcpp-generation": "sdcpp-generation";
+        "audiogen-ggml": "audiogen-ggml";
         "ggml-vla": "ggml-vla";
         "ggml-classification": "ggml-classification";
         llm: "llm";
@@ -93,6 +97,7 @@ export declare const modelInputToSrcSchema: z.ZodPipe<z.ZodUnion<readonly [z.Zod
     addon: z.ZodOptional<z.ZodUnion<readonly [z.ZodEnum<{
         embeddings: "embeddings";
         diffusion: "diffusion";
+        audiogen: "audiogen";
         tts: "tts";
         ocr: "ocr";
         vla: "vla";
@@ -106,6 +111,7 @@ export declare const modelInputToSrcSchema: z.ZodPipe<z.ZodUnion<readonly [z.Zod
         "parakeet-transcription": "parakeet-transcription";
         "ggml-ocr": "ggml-ocr";
         "sdcpp-generation": "sdcpp-generation";
+        "audiogen-ggml": "audiogen-ggml";
         "ggml-vla": "ggml-vla";
         "ggml-classification": "ggml-classification";
         llm: "llm";
@@ -126,7 +132,7 @@ export declare const modelInputToSrcSchema: z.ZodPipe<z.ZodUnion<readonly [z.Zod
     engine?: string | undefined;
     expectedSize?: number | undefined;
     sha256Checksum?: string | undefined;
-    addon?: "embeddings" | "diffusion" | "tts" | "ocr" | "vla" | "llamacpp-completion" | "whispercpp-transcription" | "bci-whispercpp-transcription" | "llamacpp-embedding" | "nmtcpp-translation" | "onnx-tts" | "tts-ggml" | "parakeet-transcription" | "ggml-ocr" | "sdcpp-generation" | "ggml-vla" | "ggml-classification" | "llm" | "whisper" | "bci" | "nmt" | "parakeet" | "classification" | "vad" | undefined;
+    addon?: "embeddings" | "diffusion" | "audiogen" | "tts" | "ocr" | "vla" | "llamacpp-completion" | "whispercpp-transcription" | "bci-whispercpp-transcription" | "llamacpp-embedding" | "nmtcpp-translation" | "onnx-tts" | "tts-ggml" | "parakeet-transcription" | "ggml-ocr" | "sdcpp-generation" | "audiogen-ggml" | "ggml-vla" | "ggml-classification" | "llm" | "whisper" | "bci" | "nmt" | "parakeet" | "classification" | "vad" | undefined;
 }>>;
 /**
  * Schema that transforms ModelSrc to its optional name
@@ -146,6 +152,7 @@ export declare const modelInputToNameSchema: z.ZodPipe<z.ZodUnion<readonly [z.Zo
     addon: z.ZodOptional<z.ZodUnion<readonly [z.ZodEnum<{
         embeddings: "embeddings";
         diffusion: "diffusion";
+        audiogen: "audiogen";
         tts: "tts";
         ocr: "ocr";
         vla: "vla";
@@ -159,6 +166,7 @@ export declare const modelInputToNameSchema: z.ZodPipe<z.ZodUnion<readonly [z.Zo
         "parakeet-transcription": "parakeet-transcription";
         "ggml-ocr": "ggml-ocr";
         "sdcpp-generation": "sdcpp-generation";
+        "audiogen-ggml": "audiogen-ggml";
         "ggml-vla": "ggml-vla";
         "ggml-classification": "ggml-classification";
         llm: "llm";
@@ -179,7 +187,7 @@ export declare const modelInputToNameSchema: z.ZodPipe<z.ZodUnion<readonly [z.Zo
     engine?: string | undefined;
     expectedSize?: number | undefined;
     sha256Checksum?: string | undefined;
-    addon?: "embeddings" | "diffusion" | "tts" | "ocr" | "vla" | "llamacpp-completion" | "whispercpp-transcription" | "bci-whispercpp-transcription" | "llamacpp-embedding" | "nmtcpp-translation" | "onnx-tts" | "tts-ggml" | "parakeet-transcription" | "ggml-ocr" | "sdcpp-generation" | "ggml-vla" | "ggml-classification" | "llm" | "whisper" | "bci" | "nmt" | "parakeet" | "classification" | "vad" | undefined;
+    addon?: "embeddings" | "diffusion" | "audiogen" | "tts" | "ocr" | "vla" | "llamacpp-completion" | "whispercpp-transcription" | "bci-whispercpp-transcription" | "llamacpp-embedding" | "nmtcpp-translation" | "onnx-tts" | "tts-ggml" | "parakeet-transcription" | "ggml-ocr" | "sdcpp-generation" | "audiogen-ggml" | "ggml-vla" | "ggml-classification" | "llm" | "whisper" | "bci" | "nmt" | "parakeet" | "classification" | "vad" | undefined;
 }>>;
 export declare function inferModelTypeFromModelSrc(modelSrc: unknown): string | undefined;
 //# sourceMappingURL=model-src-utils.d.ts.map

package/dist/schemas/model-types.d.ts

@@ -15,6 +15,7 @@ export declare const ModelType: {
     readonly parakeetTranscription: "parakeet-transcription";
     readonly ggmlOcr: "ggml-ocr";
     readonly sdcppGeneration: "sdcpp-generation";
+    readonly audiogenGgml: "audiogen-ggml";
     readonly ggmlVla: "ggml-vla";
     readonly ggmlClassification: "ggml-classification";
 };
@@ -32,6 +33,7 @@ export declare const ModelTypeAliases: {
     readonly tts: "tts-ggml";
     readonly ocr: "ggml-ocr";
     readonly diffusion: "sdcpp-generation";
+    readonly audiogen: "audiogen-ggml";
     readonly vla: "ggml-vla";
     readonly classification: "ggml-classification";
 };
@@ -63,6 +65,7 @@ export declare const PUBLIC_MODEL_TYPES: {
     readonly tts: "tts-ggml";
     readonly ocr: "ggml-ocr";
     readonly diffusion: "sdcpp-generation";
+    readonly audiogen: "audiogen-ggml";
     readonly vla: "ggml-vla";
     readonly classification: "ggml-classification";
     readonly llamacppCompletion: "llamacpp-completion";
@@ -75,6 +78,7 @@ export declare const PUBLIC_MODEL_TYPES: {
     readonly parakeetTranscription: "parakeet-transcription";
     readonly ggmlOcr: "ggml-ocr";
     readonly sdcppGeneration: "sdcpp-generation";
+    readonly audiogenGgml: "audiogen-ggml";
     readonly ggmlVla: "ggml-vla";
     readonly ggmlClassification: "ggml-classification";
 };
@@ -82,6 +86,7 @@ export declare const PUBLIC_MODEL_TYPES: {
 export declare const modelTypeInputSchema: z.ZodEnum<{
     embeddings: "embeddings";
     diffusion: "diffusion";
+    audiogen: "audiogen";
     tts: "tts";
     ocr: "ocr";
     vla: "vla";
@@ -95,6 +100,7 @@ export declare const modelTypeInputSchema: z.ZodEnum<{
     "parakeet-transcription": "parakeet-transcription";
     "ggml-ocr": "ggml-ocr";
     "sdcpp-generation": "sdcpp-generation";
+    "audiogen-ggml": "audiogen-ggml";
     "ggml-vla": "ggml-vla";
     "ggml-classification": "ggml-classification";
     llm: "llm";
@@ -108,6 +114,7 @@ export declare const modelTypeInputSchema: z.ZodEnum<{
 export declare const modelTypeSchema: z.ZodPipe<z.ZodEnum<{
     embeddings: "embeddings";
     diffusion: "diffusion";
+    audiogen: "audiogen";
     tts: "tts";
     ocr: "ocr";
     vla: "vla";
@@ -121,6 +128,7 @@ export declare const modelTypeSchema: z.ZodPipe<z.ZodEnum<{
     "parakeet-transcription": "parakeet-transcription";
     "ggml-ocr": "ggml-ocr";
     "sdcpp-generation": "sdcpp-generation";
+    "audiogen-ggml": "audiogen-ggml";
     "ggml-vla": "ggml-vla";
     "ggml-classification": "ggml-classification";
     llm: "llm";
@@ -129,7 +137,7 @@ export declare const modelTypeSchema: z.ZodPipe<z.ZodEnum<{
     nmt: "nmt";
     parakeet: "parakeet";
     classification: "classification";
-}>, z.ZodTransform<"llamacpp-completion" | "whispercpp-transcription" | "bci-whispercpp-transcription" | "llamacpp-embedding" | "nmtcpp-translation" | "onnx-tts" | "tts-ggml" | "parakeet-transcription" | "ggml-ocr" | "sdcpp-generation" | "ggml-vla" | "ggml-classification", "embeddings" | "diffusion" | "tts" | "ocr" | "vla" | "llamacpp-completion" | "whispercpp-transcription" | "bci-whispercpp-transcription" | "llamacpp-embedding" | "nmtcpp-translation" | "onnx-tts" | "tts-ggml" | "parakeet-transcription" | "ggml-ocr" | "sdcpp-generation" | "ggml-vla" | "ggml-classification" | "llm" | "whisper" | "bci" | "nmt" | "parakeet" | "classification">>;
+}>, z.ZodTransform<"llamacpp-completion" | "whispercpp-transcription" | "bci-whispercpp-transcription" | "llamacpp-embedding" | "nmtcpp-translation" | "onnx-tts" | "tts-ggml" | "parakeet-transcription" | "ggml-ocr" | "sdcpp-generation" | "audiogen-ggml" | "ggml-vla" | "ggml-classification", "embeddings" | "diffusion" | "audiogen" | "tts" | "ocr" | "vla" | "llamacpp-completion" | "whispercpp-transcription" | "bci-whispercpp-transcription" | "llamacpp-embedding" | "nmtcpp-translation" | "onnx-tts" | "tts-ggml" | "parakeet-transcription" | "ggml-ocr" | "sdcpp-generation" | "audiogen-ggml" | "ggml-vla" | "ggml-classification" | "llm" | "whisper" | "bci" | "nmt" | "parakeet" | "classification">>;
 /**
  * Normalize model type input to canonical form.
  * Custom plugin types (not in built-ins) are returned as-is.
@@ -235,6 +243,16 @@ export declare const diffusionModelTypeSchema: z.ZodEnum<{
     "sdcpp-generation": "sdcpp-generation";
 }>;
 export type DiffusionModelTypeInput = z.infer<typeof diffusionModelTypeSchema>;
+/**
+ * AudioGen model type schema.
+ * - Alias: `"audiogen"` → resolves to `"audiogen-ggml"`
+ * - Canonical: `"audiogen-ggml"`
+ */
+export declare const audioGenModelTypeSchema: z.ZodEnum<{
+    audiogen: "audiogen";
+    "audiogen-ggml": "audiogen-ggml";
+}>;
+export type AudioGenModelTypeInput = z.infer<typeof audioGenModelTypeSchema>;
 /**
  * VLA (vision-language-action) model type schema.
  * - Alias: `"vla"` → resolves to `"ggml-vla"`

package/dist/schemas/model-types.js

@@ -16,6 +16,7 @@ export const ModelType = {
     parakeetTranscription: 'parakeet-transcription',
     ggmlOcr: 'ggml-ocr',
     sdcppGeneration: 'sdcpp-generation',
+    audiogenGgml: 'audiogen-ggml',
     ggmlVla: 'ggml-vla',
     ggmlClassification: 'ggml-classification'
 };
@@ -30,6 +31,7 @@ const AliasKeys = {
     tts: 'tts',
     ocr: 'ocr',
     diffusion: 'diffusion',
+    audiogen: 'audiogen',
     vla: 'vla',
     classification: 'classification'
 };
@@ -48,6 +50,7 @@ export const ModelTypeAliases = {
     [AliasKeys.tts]: ModelType.ttsGgml,
     [AliasKeys.ocr]: ModelType.ggmlOcr,
     [AliasKeys.diffusion]: ModelType.sdcppGeneration,
+    [AliasKeys.audiogen]: ModelType.audiogenGgml,
     [AliasKeys.vla]: ModelType.ggmlVla,
     [AliasKeys.classification]: ModelType.ggmlClassification
 };
@@ -194,6 +197,14 @@ export const ocrModelTypeSchema = modelTypeInputSchema
 export const diffusionModelTypeSchema = modelTypeInputSchema
     .extract([AliasKeys.diffusion, ModelType.sdcppGeneration])
     .describe('Diffusion model type: "diffusion" (alias) or "sdcpp-generation" (canonical)');
+/**
+ * AudioGen model type schema.
+ * - Alias: `"audiogen"` → resolves to `"audiogen-ggml"`
+ * - Canonical: `"audiogen-ggml"`
+ */
+export const audioGenModelTypeSchema = modelTypeInputSchema
+    .extract([AliasKeys.audiogen, ModelType.audiogenGgml])
+    .describe('AudioGen model type: "audiogen" (alias) or "audiogen-ggml" (canonical)');
 /**
  * VLA (vision-language-action) model type schema.
  * - Alias: `"vla"` → resolves to `"ggml-vla"`

package/dist/schemas/ocr.d.ts

@@ -33,6 +33,7 @@ export declare const ocrConfigSchema: z.ZodObject<{
         addon: z.ZodOptional<z.ZodUnion<readonly [z.ZodEnum<{
             embeddings: "embeddings";
             diffusion: "diffusion";
+            audiogen: "audiogen";
             tts: "tts";
             ocr: "ocr";
             vla: "vla";
@@ -46,6 +47,7 @@ export declare const ocrConfigSchema: z.ZodObject<{
             "parakeet-transcription": "parakeet-transcription";
             "ggml-ocr": "ggml-ocr";
             "sdcpp-generation": "sdcpp-generation";
+            "audiogen-ggml": "audiogen-ggml";
             "ggml-vla": "ggml-vla";
             "ggml-classification": "ggml-classification";
             llm: "llm";

package/dist/schemas/plugin.d.ts

@@ -1,5 +1,5 @@
 import { z } from 'zod';
-import type { ModelSrcInput } from './model-src-utils';
+import type { ModelSrcInput } from './model-src-utils.js';
 /**
  * Granularity at which the addon can cancel.
  *  - `"request"` — addon cancels a specific in-flight `requestId`.
@@ -286,10 +286,15 @@ export declare const PLUGIN_OCR: "@qvac/sdk/ggml-ocr/plugin";
  */
 export declare const PLUGIN_DIFFUSION: "@qvac/sdk/sdcpp-generation/plugin";
 /**
- * Vision-Language-Action plugin (ggml). Supports SmolVLA and π₀.₅ (pi05),
- * dispatched on the GGUF `general.architecture` key.
+ * Music and audio generation plugin (GGML / ACE-Step).
+ * Provides: text-conditioned PCM audio generation.
+ */
+export declare const PLUGIN_AUDIOGEN: "@qvac/sdk/audiogen-ggml/plugin";
+/**
+ * Vision-Language-Action plugin (ggml). Supports SmolVLA, π₀.₅ (pi05), and
+ * GR00T, dispatched on the GGUF `general.architecture` key.
  * Provides: robot-action inference from one or more camera frames (2 for
- * SmolVLA, 3 for π₀.₅) and a natural-language instruction.
+ * SmolVLA and GR00T, 3 for π₀.₅) and a natural-language instruction.
  */
 export declare const PLUGIN_VLA: "@qvac/sdk/ggml-vla/plugin";
 /**
@@ -306,17 +311,19 @@ export declare const PLUGIN_CLASSIFICATION: "@qvac/sdk/ggml-classification/plugi
  *   plugins: [...SDK_DEFAULT_PLUGINS, myCustomPlugin],
  * };
  */
-export declare const SDK_DEFAULT_PLUGINS: readonly ["@qvac/sdk/llamacpp-completion/plugin", "@qvac/sdk/llamacpp-embedding/plugin", "@qvac/sdk/whispercpp-transcription/plugin", "@qvac/sdk/bci-whispercpp-transcription/plugin", "@qvac/sdk/parakeet-transcription/plugin", "@qvac/sdk/nmtcpp-translation/plugin", "@qvac/sdk/tts-ggml/plugin", "@qvac/sdk/ggml-ocr/plugin", "@qvac/sdk/sdcpp-generation/plugin", "@qvac/sdk/ggml-vla/plugin", "@qvac/sdk/ggml-classification/plugin"];
+export declare const SDK_DEFAULT_PLUGINS: readonly ["@qvac/sdk/llamacpp-completion/plugin", "@qvac/sdk/llamacpp-embedding/plugin", "@qvac/sdk/whispercpp-transcription/plugin", "@qvac/sdk/bci-whispercpp-transcription/plugin", "@qvac/sdk/parakeet-transcription/plugin", "@qvac/sdk/nmtcpp-translation/plugin", "@qvac/sdk/tts-ggml/plugin", "@qvac/sdk/ggml-ocr/plugin", "@qvac/sdk/sdcpp-generation/plugin", "@qvac/sdk/audiogen-ggml/plugin", "@qvac/sdk/ggml-vla/plugin", "@qvac/sdk/ggml-classification/plugin"];
 export type BuiltinPlugin = (typeof SDK_DEFAULT_PLUGINS)[number];
 /** Native addon package for LLM completion (llama.cpp) */
 export declare const ADDON_LLM: "@qvac/llm-llamacpp";
 /** Native addon package for text embeddings (llama.cpp) */
 export declare const ADDON_EMBEDDING: "@qvac/embed-llamacpp";
-/** Native addon package for Whisper transcription (whisper.cpp) */
+/** Unified native addon package for Whisper and Parakeet transcription */
+export declare const ADDON_ASR: "@qvac/asr-ggml";
+/** Legacy native addon package for Whisper transcription (whisper.cpp) */
 export declare const ADDON_WHISPER: "@qvac/transcription-whispercpp";
 /** Native addon package for BCI neural-signal transcription (whisper.cpp) */
 export declare const ADDON_BCI: "@qvac/bci-whispercpp";
-/** Native addon package for Parakeet transcription (ONNX) */
+/** Legacy native addon package for Parakeet transcription */
 export declare const ADDON_PARAKEET: "@qvac/transcription-parakeet";
 /** Native addon package for NMT translation (nmt.cpp) */
 export declare const ADDON_NMT: "@qvac/translation-nmtcpp";
@@ -326,7 +333,9 @@ export declare const ADDON_TTS: "@qvac/tts-ggml";
 export declare const ADDON_OCR: "@qvac/ocr-ggml";
 /** Native addon package for image generation (stable-diffusion.cpp) */
 export declare const ADDON_DIFFUSION: "@qvac/diffusion-cpp";
-/** Native addon package for vision-language-action inference (SmolVLA / π₀.₅ on ggml) */
+/** Native addon package for music and audio generation (GGML / ACE-Step) */
+export declare const ADDON_AUDIOGEN: "@qvac/audiogen-ggml";
+/** Native addon package for vision-language-action inference (SmolVLA / π₀.₅ / GR00T on ggml) */
 export declare const ADDON_VLA: "@qvac/vla-ggml";
 /** Native addon package for image classification (GGML / MobileNetV3) */
 export declare const ADDON_CLASSIFICATION: "@qvac/classification-ggml";

package/dist/schemas/plugin.js

@@ -172,10 +172,15 @@ export const PLUGIN_OCR = '@qvac/sdk/ggml-ocr/plugin';
  */
 export const PLUGIN_DIFFUSION = '@qvac/sdk/sdcpp-generation/plugin';
 /**
- * Vision-Language-Action plugin (ggml). Supports SmolVLA and π₀.₅ (pi05),
- * dispatched on the GGUF `general.architecture` key.
+ * Music and audio generation plugin (GGML / ACE-Step).
+ * Provides: text-conditioned PCM audio generation.
+ */
+export const PLUGIN_AUDIOGEN = '@qvac/sdk/audiogen-ggml/plugin';
+/**
+ * Vision-Language-Action plugin (ggml). Supports SmolVLA, π₀.₅ (pi05), and
+ * GR00T, dispatched on the GGUF `general.architecture` key.
  * Provides: robot-action inference from one or more camera frames (2 for
- * SmolVLA, 3 for π₀.₅) and a natural-language instruction.
+ * SmolVLA and GR00T, 3 for π₀.₅) and a natural-language instruction.
  */
 export const PLUGIN_VLA = '@qvac/sdk/ggml-vla/plugin';
 /**
@@ -202,6 +207,7 @@ export const SDK_DEFAULT_PLUGINS = [
     PLUGIN_TTS,
     PLUGIN_OCR,
     PLUGIN_DIFFUSION,
+    PLUGIN_AUDIOGEN,
     PLUGIN_VLA,
     PLUGIN_CLASSIFICATION
 ];
@@ -212,11 +218,13 @@ export const SDK_DEFAULT_PLUGINS = [
 export const ADDON_LLM = '@qvac/llm-llamacpp';
 /** Native addon package for text embeddings (llama.cpp) */
 export const ADDON_EMBEDDING = '@qvac/embed-llamacpp';
-/** Native addon package for Whisper transcription (whisper.cpp) */
+/** Unified native addon package for Whisper and Parakeet transcription */
+export const ADDON_ASR = '@qvac/asr-ggml';
+/** Legacy native addon package for Whisper transcription (whisper.cpp) */
 export const ADDON_WHISPER = '@qvac/transcription-whispercpp';
 /** Native addon package for BCI neural-signal transcription (whisper.cpp) */
 export const ADDON_BCI = '@qvac/bci-whispercpp';
-/** Native addon package for Parakeet transcription (ONNX) */
+/** Legacy native addon package for Parakeet transcription */
 export const ADDON_PARAKEET = '@qvac/transcription-parakeet';
 /** Native addon package for NMT translation (nmt.cpp) */
 export const ADDON_NMT = '@qvac/translation-nmtcpp';
@@ -226,7 +234,9 @@ export const ADDON_TTS = '@qvac/tts-ggml';
 export const ADDON_OCR = '@qvac/ocr-ggml';
 /** Native addon package for image generation (stable-diffusion.cpp) */
 export const ADDON_DIFFUSION = '@qvac/diffusion-cpp';
-/** Native addon package for vision-language-action inference (SmolVLA / π₀.₅ on ggml) */
+/** Native addon package for music and audio generation (GGML / ACE-Step) */
+export const ADDON_AUDIOGEN = '@qvac/audiogen-ggml';
+/** Native addon package for vision-language-action inference (SmolVLA / π₀.₅ / GR00T on ggml) */
 export const ADDON_VLA = '@qvac/vla-ggml';
 /** Native addon package for image classification (GGML / MobileNetV3) */
 export const ADDON_CLASSIFICATION = '@qvac/classification-ggml';