@nebutra/audio-pipeline 0.1.0

package/.turbo/turbo-build.log

@@ -0,0 +1,22 @@
+
+> @nebutra/audio-pipeline@0.1.0 build /home/runner/work/Nebutra-Sailor/Nebutra-Sailor/packages/ai/audio-pipeline
+> tsup
+
+CLI Building entry: src/cli.ts, src/index.ts
+CLI Using tsconfig: tsconfig.json
+CLI tsup v8.5.1
+CLI Using tsup config: /home/runner/work/Nebutra-Sailor/Nebutra-Sailor/packages/ai/audio-pipeline/tsup.config.ts
+CLI Target: es2022
+CLI Cleaning output folder
+ESM Build start
+ESM dist/cli.js                1.32 KB
+ESM dist/index.js              186.00 B
+ESM dist/chunk-32LVUMTP.js     4.46 KB
+ESM dist/cli.js.map            2.35 KB
+ESM dist/index.js.map          71.00 B
+ESM dist/chunk-32LVUMTP.js.map 8.83 KB
+ESM ⚡️ Build success in 91ms
+DTS Build start
+DTS ⚡️ Build success in 11357ms
+DTS dist/cli.d.ts   13.00 B
+DTS dist/index.d.ts 1.47 KB

package/.turbo/turbo-test.log

@@ -0,0 +1,14 @@
+
+> @nebutra/audio-pipeline@0.1.0 test /home/runner/work/Nebutra-Sailor/Nebutra-Sailor/packages/ai/audio-pipeline
+> vitest run
+
+
+ RUN  v4.1.4 /home/runner/work/Nebutra-Sailor/Nebutra-Sailor/packages/ai/audio-pipeline
+
+ ✓ src/index.test.ts (2 tests) 163ms
+
+ Test Files  1 passed (1)
+      Tests  2 passed (2)
+   Start at  05:41:39
+   Duration  1.49s (transform 462ms, setup 0ms, import 588ms, tests 163ms, environment 0ms)
+

package/.turbo/turbo-typecheck.log

@@ -0,0 +1,4 @@
+
+> @nebutra/audio-pipeline@0.1.0 typecheck /home/runner/work/Nebutra-Sailor/Nebutra-Sailor/packages/ai/audio-pipeline
+> tsc --noEmit
+

package/README.md

@@ -0,0 +1,19 @@
+# @nebutra/audio-pipeline
+
+Status: WIP — Not yet integrated into any production app.
+
+`@nebutra/audio-pipeline` is the BrandContext-first audio capability surface.
+The zero-config path writes short valid WAV files with license metadata so
+doctor, debug, and examples are executable without model credentials.
+
+It does not own prompt orchestration, Thread/Turn/Item state, model provider
+routing, or approval lifecycle.
+
+## Commands
+
+```bash
+pnpm audio:doctor
+pnpm audio:debug
+pnpm audio:license <asset>
+pnpm audio:loudness <asset>
+```

package/dist/chunk-32LVUMTP.js

@@ -0,0 +1,141 @@
+// src/index.ts
+import { mkdir, writeFile } from "fs/promises";
+import { dirname, join } from "path";
+import { appendCapabilityDebug, readCapabilityDebug } from "@nebutra/capability-kit/debug";
+import { CapabilityError } from "@nebutra/errors";
+import {
+  assetId,
+  requireBrandContext
+} from "@nebutra/generation-context";
+async function readAudioDebug(root = process.cwd(), limit = 10) {
+  return readCapabilityDebug("audio-pipeline", { root, limit });
+}
+function intentLabel(intent) {
+  switch (intent.type) {
+    case "bgm":
+      return intent.mood;
+    case "sfx":
+      return intent.description;
+    case "song":
+      return intent.lyricsPrompt;
+  }
+}
+function durationFor(intent) {
+  if (intent.type === "sfx") return Math.max(1, Math.min(intent.durationS ?? 3, 8));
+  return Math.max(1, Math.min(intent.durationS, 12));
+}
+function createToneWav(durationS, frequency = 440, sampleRate = 24e3) {
+  const samples = Math.max(1, Math.floor(durationS * sampleRate));
+  const dataBytes = samples * 2;
+  const buffer = Buffer.alloc(44 + dataBytes);
+  buffer.write("RIFF", 0);
+  buffer.writeUInt32LE(36 + dataBytes, 4);
+  buffer.write("WAVE", 8);
+  buffer.write("fmt ", 12);
+  buffer.writeUInt32LE(16, 16);
+  buffer.writeUInt16LE(1, 20);
+  buffer.writeUInt16LE(1, 22);
+  buffer.writeUInt32LE(sampleRate, 24);
+  buffer.writeUInt32LE(sampleRate * 2, 28);
+  buffer.writeUInt16LE(2, 32);
+  buffer.writeUInt16LE(16, 34);
+  buffer.write("data", 36);
+  buffer.writeUInt32LE(dataBytes, 40);
+  for (let index = 0; index < samples; index += 1) {
+    const fade = Math.min(index / 1200, (samples - index) / 1200, 1);
+    const sample = Math.round(
+      Math.sin(2 * Math.PI * frequency * index / sampleRate) * 9e3 * fade
+    );
+    buffer.writeInt16LE(sample, 44 + index * 2);
+  }
+  return buffer;
+}
+function frequencyFor(brand, intent) {
+  const base = brand.toneKeywords.join("").length + intentLabel(intent).length;
+  return 220 + base % 9 * 55;
+}
+function licenseFor(requireCommercial, provider) {
+  if (requireCommercial && provider !== "tone-local") {
+    return {
+      status: "unknown",
+      source: provider,
+      suggestion: "Verify the provider plan and model terms before commercial use."
+    };
+  }
+  return { status: "commercial-ok", source: "deterministic local renderer" };
+}
+var AudioPipeline = class {
+  #root;
+  #provider;
+  constructor(options = {}) {
+    this.#root = options.root ?? process.cwd();
+    this.#provider = options.provider ?? "tone-local";
+  }
+  async generate(intent, brandInput, requireCommercial = true) {
+    const brand = requireBrandContext(brandInput, "audio-pipeline");
+    if (requireCommercial && this.#provider !== "tone-local") {
+      throw new CapabilityError(
+        "audio-pipeline",
+        "Selected audio provider lacks verified license",
+        {
+          suggestion: "Use tone-local or wire a provider with commercial license metadata.",
+          statusCode: 409
+        }
+      );
+    }
+    const durationS = durationFor(intent);
+    const id = assetId("audio", `${intent.type}_${intentLabel(intent)}`);
+    const path = join(this.#root, ".nebutra", "generated", "audio-pipeline", `${id}.wav`);
+    await mkdir(dirname(path), { recursive: true });
+    await writeFile(path, createToneWav(durationS, frequencyFor(brand, intent)));
+    const asset = {
+      id,
+      tenantId: brand.tenantId,
+      kind: "audio",
+      path,
+      brandId: brand.brandId,
+      provider: this.#provider,
+      model: "brand-context-tone-v1",
+      createdAt: (/* @__PURE__ */ new Date()).toISOString(),
+      license: licenseFor(requireCommercial, this.#provider),
+      durationS,
+      format: "wav",
+      loudnessLufs: -18,
+      metadata: { intent, brandSource: brand.sourcePath }
+    };
+    await appendCapabilityDebug(
+      "audio-pipeline",
+      { type: "generate", asset },
+      { root: this.#root }
+    );
+    return asset;
+  }
+  async doctor() {
+    return [
+      { provider: "tone-local", ok: true },
+      {
+        provider: "local-model",
+        ok: Boolean(process.env.AUDIO_LOCAL_MODEL_PATH),
+        suggestion: "Set AUDIO_LOCAL_MODEL_PATH to enable local model generation."
+      },
+      {
+        provider: "remote",
+        ok: Boolean(process.env.AUDIO_REMOTE_API_KEY),
+        suggestion: "Set AUDIO_REMOTE_API_KEY to enable remote audio fallback."
+      }
+    ];
+  }
+  async license(asset) {
+    return asset.license;
+  }
+  async loudness(asset) {
+    return { lufs: asset.loudnessLufs };
+  }
+};
+
+export {
+  readAudioDebug,
+  createToneWav,
+  AudioPipeline
+};
+//# sourceMappingURL=chunk-32LVUMTP.js.map

package/dist/chunk-32LVUMTP.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/index.ts"],"sourcesContent":["import { mkdir, writeFile } from \"node:fs/promises\";\nimport { dirname, join } from \"node:path\";\nimport { appendCapabilityDebug, readCapabilityDebug } from \"@nebutra/capability-kit/debug\";\nimport { CapabilityError } from \"@nebutra/errors\";\nimport {\n  assetId,\n  type BrandContext,\n  type GeneratedAsset,\n  type LicenseMetadata,\n  requireBrandContext,\n} from \"@nebutra/generation-context\";\n\nexport type AudioIntent =\n  | { type: \"bgm\"; durationS: number; mood: string; bpm?: number }\n  | { type: \"sfx\"; description: string; durationS?: number }\n  | { type: \"song\"; lyricsPrompt: string; durationS: number };\n\nexport interface AudioAsset extends GeneratedAsset {\n  readonly kind: \"audio\";\n  readonly durationS: number;\n  readonly format: \"wav\";\n  readonly loudnessLufs: number;\n}\n\nexport interface AudioHealth {\n  readonly provider: string;\n  readonly ok: boolean;\n  readonly suggestion?: string;\n}\n\nexport interface AudioPipelineOptions {\n  readonly root?: string;\n  readonly provider?: \"tone-local\" | \"local-model\" | \"remote\";\n}\n\nexport async function readAudioDebug(root = process.cwd(), limit = 10): Promise<unknown[]> {\n  return readCapabilityDebug(\"audio-pipeline\", { root, limit });\n}\n\nfunction intentLabel(intent: AudioIntent): string {\n  switch (intent.type) {\n    case \"bgm\":\n      return intent.mood;\n    case \"sfx\":\n      return intent.description;\n    case \"song\":\n      return intent.lyricsPrompt;\n  }\n}\n\nfunction durationFor(intent: AudioIntent): number {\n  if (intent.type === \"sfx\") return Math.max(1, Math.min(intent.durationS ?? 3, 8));\n  return Math.max(1, Math.min(intent.durationS, 12));\n}\n\nexport function createToneWav(durationS: number, frequency = 440, sampleRate = 24_000): Buffer {\n  const samples = Math.max(1, Math.floor(durationS * sampleRate));\n  const dataBytes = samples * 2;\n  const buffer = Buffer.alloc(44 + dataBytes);\n  buffer.write(\"RIFF\", 0);\n  buffer.writeUInt32LE(36 + dataBytes, 4);\n  buffer.write(\"WAVE\", 8);\n  buffer.write(\"fmt \", 12);\n  buffer.writeUInt32LE(16, 16);\n  buffer.writeUInt16LE(1, 20);\n  buffer.writeUInt16LE(1, 22);\n  buffer.writeUInt32LE(sampleRate, 24);\n  buffer.writeUInt32LE(sampleRate * 2, 28);\n  buffer.writeUInt16LE(2, 32);\n  buffer.writeUInt16LE(16, 34);\n  buffer.write(\"data\", 36);\n  buffer.writeUInt32LE(dataBytes, 40);\n  for (let index = 0; index < samples; index += 1) {\n    const fade = Math.min(index / 1200, (samples - index) / 1200, 1);\n    const sample = Math.round(\n      Math.sin((2 * Math.PI * frequency * index) / sampleRate) * 9000 * fade,\n    );\n    buffer.writeInt16LE(sample, 44 + index * 2);\n  }\n  return buffer;\n}\n\nfunction frequencyFor(brand: BrandContext, intent: AudioIntent): number {\n  const base = brand.toneKeywords.join(\"\").length + intentLabel(intent).length;\n  return 220 + (base % 9) * 55;\n}\n\nfunction licenseFor(requireCommercial: boolean, provider: string): LicenseMetadata {\n  if (requireCommercial && provider !== \"tone-local\") {\n    return {\n      status: \"unknown\",\n      source: provider,\n      suggestion: \"Verify the provider plan and model terms before commercial use.\",\n    };\n  }\n  return { status: \"commercial-ok\", source: \"deterministic local renderer\" };\n}\n\nexport class AudioPipeline {\n  readonly #root: string;\n  readonly #provider: \"tone-local\" | \"local-model\" | \"remote\";\n\n  constructor(options: AudioPipelineOptions = {}) {\n    this.#root = options.root ?? process.cwd();\n    this.#provider = options.provider ?? \"tone-local\";\n  }\n\n  async generate(\n    intent: AudioIntent,\n    brandInput: BrandContext | undefined,\n    requireCommercial = true,\n  ): Promise<AudioAsset> {\n    const brand = requireBrandContext(brandInput, \"audio-pipeline\");\n    if (requireCommercial && this.#provider !== \"tone-local\") {\n      throw new CapabilityError(\n        \"audio-pipeline\",\n        \"Selected audio provider lacks verified license\",\n        {\n          suggestion: \"Use tone-local or wire a provider with commercial license metadata.\",\n          statusCode: 409,\n        },\n      );\n    }\n    const durationS = durationFor(intent);\n    const id = assetId(\"audio\", `${intent.type}_${intentLabel(intent)}`);\n    const path = join(this.#root, \".nebutra\", \"generated\", \"audio-pipeline\", `${id}.wav`);\n    await mkdir(dirname(path), { recursive: true });\n    await writeFile(path, createToneWav(durationS, frequencyFor(brand, intent)));\n\n    const asset: AudioAsset = {\n      id,\n      tenantId: brand.tenantId,\n      kind: \"audio\",\n      path,\n      brandId: brand.brandId,\n      provider: this.#provider,\n      model: \"brand-context-tone-v1\",\n      createdAt: new Date().toISOString(),\n      license: licenseFor(requireCommercial, this.#provider),\n      durationS,\n      format: \"wav\",\n      loudnessLufs: -18,\n      metadata: { intent, brandSource: brand.sourcePath },\n    };\n    await appendCapabilityDebug(\n      \"audio-pipeline\",\n      { type: \"generate\", asset },\n      { root: this.#root },\n    );\n    return asset;\n  }\n\n  async doctor(): Promise<AudioHealth[]> {\n    return [\n      { provider: \"tone-local\", ok: true },\n      {\n        provider: \"local-model\",\n        ok: Boolean(process.env.AUDIO_LOCAL_MODEL_PATH),\n        suggestion: \"Set AUDIO_LOCAL_MODEL_PATH to enable local model generation.\",\n      },\n      {\n        provider: \"remote\",\n        ok: Boolean(process.env.AUDIO_REMOTE_API_KEY),\n        suggestion: \"Set AUDIO_REMOTE_API_KEY to enable remote audio fallback.\",\n      },\n    ];\n  }\n\n  async license(asset: Pick<AudioAsset, \"license\">): Promise<LicenseMetadata> {\n    return asset.license;\n  }\n\n  async loudness(asset: Pick<AudioAsset, \"loudnessLufs\">): Promise<{ lufs: number }> {\n    return { lufs: asset.loudnessLufs };\n  }\n}\n"],"mappings":";AAAA,SAAS,OAAO,iBAAiB;AACjC,SAAS,SAAS,YAAY;AAC9B,SAAS,uBAAuB,2BAA2B;AAC3D,SAAS,uBAAuB;AAChC;AAAA,EACE;AAAA,EAIA;AAAA,OACK;AAyBP,eAAsB,eAAe,OAAO,QAAQ,IAAI,GAAG,QAAQ,IAAwB;AACzF,SAAO,oBAAoB,kBAAkB,EAAE,MAAM,MAAM,CAAC;AAC9D;AAEA,SAAS,YAAY,QAA6B;AAChD,UAAQ,OAAO,MAAM;AAAA,IACnB,KAAK;AACH,aAAO,OAAO;AAAA,IAChB,KAAK;AACH,aAAO,OAAO;AAAA,IAChB,KAAK;AACH,aAAO,OAAO;AAAA,EAClB;AACF;AAEA,SAAS,YAAY,QAA6B;AAChD,MAAI,OAAO,SAAS,MAAO,QAAO,KAAK,IAAI,GAAG,KAAK,IAAI,OAAO,aAAa,GAAG,CAAC,CAAC;AAChF,SAAO,KAAK,IAAI,GAAG,KAAK,IAAI,OAAO,WAAW,EAAE,CAAC;AACnD;AAEO,SAAS,cAAc,WAAmB,YAAY,KAAK,aAAa,MAAgB;AAC7F,QAAM,UAAU,KAAK,IAAI,GAAG,KAAK,MAAM,YAAY,UAAU,CAAC;AAC9D,QAAM,YAAY,UAAU;AAC5B,QAAM,SAAS,OAAO,MAAM,KAAK,SAAS;AAC1C,SAAO,MAAM,QAAQ,CAAC;AACtB,SAAO,cAAc,KAAK,WAAW,CAAC;AACtC,SAAO,MAAM,QAAQ,CAAC;AACtB,SAAO,MAAM,QAAQ,EAAE;AACvB,SAAO,cAAc,IAAI,EAAE;AAC3B,SAAO,cAAc,GAAG,EAAE;AAC1B,SAAO,cAAc,GAAG,EAAE;AAC1B,SAAO,cAAc,YAAY,EAAE;AACnC,SAAO,cAAc,aAAa,GAAG,EAAE;AACvC,SAAO,cAAc,GAAG,EAAE;AAC1B,SAAO,cAAc,IAAI,EAAE;AAC3B,SAAO,MAAM,QAAQ,EAAE;AACvB,SAAO,cAAc,WAAW,EAAE;AAClC,WAAS,QAAQ,GAAG,QAAQ,SAAS,SAAS,GAAG;AAC/C,UAAM,OAAO,KAAK,IAAI,QAAQ,OAAO,UAAU,SAAS,MAAM,CAAC;AAC/D,UAAM,SAAS,KAAK;AAAA,MAClB,KAAK,IAAK,IAAI,KAAK,KAAK,YAAY,QAAS,UAAU,IAAI,MAAO;AAAA,IACpE;AACA,WAAO,aAAa,QAAQ,KAAK,QAAQ,CAAC;AAAA,EAC5C;AACA,SAAO;AACT;AAEA,SAAS,aAAa,OAAqB,QAA6B;AACtE,QAAM,OAAO,MAAM,aAAa,KAAK,EAAE,EAAE,SAAS,YAAY,MAAM,EAAE;AACtE,SAAO,MAAO,OAAO,IAAK;AAC5B;AAEA,SAAS,WAAW,mBAA4B,UAAmC;AACjF,MAAI,qBAAqB,aAAa,cAAc;AAClD,WAAO;AAAA,MACL,QAAQ;AAAA,MACR,QAAQ;AAAA,MACR,YAAY;AAAA,IACd;AAAA,EACF;AACA,SAAO,EAAE,QAAQ,iBAAiB,QAAQ,+BAA+B;AAC3E;AAEO,IAAM,gBAAN,MAAoB;AAAA,EAChB;AAAA,EACA;AAAA,EAET,YAAY,UAAgC,CAAC,GAAG;AAC9C,SAAK,QAAQ,QAAQ,QAAQ,QAAQ,IAAI;AACzC,SAAK,YAAY,QAAQ,YAAY;AAAA,EACvC;AAAA,EAEA,MAAM,SACJ,QACA,YACA,oBAAoB,MACC;AACrB,UAAM,QAAQ,oBAAoB,YAAY,gBAAgB;AAC9D,QAAI,qBAAqB,KAAK,cAAc,cAAc;AACxD,YAAM,IAAI;AAAA,QACR;AAAA,QACA;AAAA,QACA;AAAA,UACE,YAAY;AAAA,UACZ,YAAY;AAAA,QACd;AAAA,MACF;AAAA,IACF;AACA,UAAM,YAAY,YAAY,MAAM;AACpC,UAAM,KAAK,QAAQ,SAAS,GAAG,OAAO,IAAI,IAAI,YAAY,MAAM,CAAC,EAAE;AACnE,UAAM,OAAO,KAAK,KAAK,OAAO,YAAY,aAAa,kBAAkB,GAAG,EAAE,MAAM;AACpF,UAAM,MAAM,QAAQ,IAAI,GAAG,EAAE,WAAW,KAAK,CAAC;AAC9C,UAAM,UAAU,MAAM,cAAc,WAAW,aAAa,OAAO,MAAM,CAAC,CAAC;AAE3E,UAAM,QAAoB;AAAA,MACxB;AAAA,MACA,UAAU,MAAM;AAAA,MAChB,MAAM;AAAA,MACN;AAAA,MACA,SAAS,MAAM;AAAA,MACf,UAAU,KAAK;AAAA,MACf,OAAO;AAAA,MACP,YAAW,oBAAI,KAAK,GAAE,YAAY;AAAA,MAClC,SAAS,WAAW,mBAAmB,KAAK,SAAS;AAAA,MACrD;AAAA,MACA,QAAQ;AAAA,MACR,cAAc;AAAA,MACd,UAAU,EAAE,QAAQ,aAAa,MAAM,WAAW;AAAA,IACpD;AACA,UAAM;AAAA,MACJ;AAAA,MACA,EAAE,MAAM,YAAY,MAAM;AAAA,MAC1B,EAAE,MAAM,KAAK,MAAM;AAAA,IACrB;AACA,WAAO;AAAA,EACT;AAAA,EAEA,MAAM,SAAiC;AACrC,WAAO;AAAA,MACL,EAAE,UAAU,cAAc,IAAI,KAAK;AAAA,MACnC;AAAA,QACE,UAAU;AAAA,QACV,IAAI,QAAQ,QAAQ,IAAI,sBAAsB;AAAA,QAC9C,YAAY;AAAA,MACd;AAAA,MACA;AAAA,QACE,UAAU;AAAA,QACV,IAAI,QAAQ,QAAQ,IAAI,oBAAoB;AAAA,QAC5C,YAAY;AAAA,MACd;AAAA,IACF;AAAA,EACF;AAAA,EAEA,MAAM,QAAQ,OAA8D;AAC1E,WAAO,MAAM;AAAA,EACf;AAAA,EAEA,MAAM,SAAS,OAAoE;AACjF,WAAO,EAAE,MAAM,MAAM,aAAa;AAAA,EACpC;AACF;","names":[]}

package/dist/cli.d.ts

@@ -0,0 +1,2 @@
+
+export {  }

package/dist/cli.js

@@ -0,0 +1,45 @@
+import {
+  AudioPipeline,
+  readAudioDebug
+} from "./chunk-32LVUMTP.js";
+
+// src/cli.ts
+import { createDemoBrandContext } from "@nebutra/generation-context";
+var command = process.argv[2] ?? "doctor";
+var pipeline = new AudioPipeline();
+if (command === "doctor") {
+  process.stdout.write(
+    `${JSON.stringify({ capability: "audio-pipeline", results: await pipeline.doctor() }, null, 2)}
+`
+  );
+} else if (command === "debug") {
+  process.stdout.write(
+    `${JSON.stringify({ capability: "audio-pipeline", entries: await readAudioDebug() }, null, 2)}
+`
+  );
+} else if (command === "license") {
+  const asset = await pipeline.generate(
+    { type: "bgm", durationS: 3, mood: "uplifting technical" },
+    createDemoBrandContext(),
+    true
+  );
+  process.stdout.write(
+    `${JSON.stringify({ capability: "audio-pipeline", license: await pipeline.license(asset) }, null, 2)}
+`
+  );
+} else if (command === "loudness") {
+  const asset = await pipeline.generate(
+    { type: "sfx", description: "keyboard typing", durationS: 1 },
+    createDemoBrandContext(),
+    true
+  );
+  process.stdout.write(
+    `${JSON.stringify({ capability: "audio-pipeline", loudness: await pipeline.loudness(asset) }, null, 2)}
+`
+  );
+} else {
+  process.stderr.write(`Unknown audio-pipeline command: ${command}
+`);
+  process.exitCode = 1;
+}
+//# sourceMappingURL=cli.js.map

package/dist/cli.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/cli.ts"],"sourcesContent":["import { createDemoBrandContext } from \"@nebutra/generation-context\";\nimport { AudioPipeline, readAudioDebug } from \"./index\";\n\nconst command = process.argv[2] ?? \"doctor\";\nconst pipeline = new AudioPipeline();\n\nif (command === \"doctor\") {\n  process.stdout.write(\n    `${JSON.stringify({ capability: \"audio-pipeline\", results: await pipeline.doctor() }, null, 2)}\\n`,\n  );\n} else if (command === \"debug\") {\n  process.stdout.write(\n    `${JSON.stringify({ capability: \"audio-pipeline\", entries: await readAudioDebug() }, null, 2)}\\n`,\n  );\n} else if (command === \"license\") {\n  const asset = await pipeline.generate(\n    { type: \"bgm\", durationS: 3, mood: \"uplifting technical\" },\n    createDemoBrandContext(),\n    true,\n  );\n  process.stdout.write(\n    `${JSON.stringify({ capability: \"audio-pipeline\", license: await pipeline.license(asset) }, null, 2)}\\n`,\n  );\n} else if (command === \"loudness\") {\n  const asset = await pipeline.generate(\n    { type: \"sfx\", description: \"keyboard typing\", durationS: 1 },\n    createDemoBrandContext(),\n    true,\n  );\n  process.stdout.write(\n    `${JSON.stringify({ capability: \"audio-pipeline\", loudness: await pipeline.loudness(asset) }, null, 2)}\\n`,\n  );\n} else {\n  process.stderr.write(`Unknown audio-pipeline command: ${command}\\n`);\n  process.exitCode = 1;\n}\n"],"mappings":";;;;;;AAAA,SAAS,8BAA8B;AAGvC,IAAM,UAAU,QAAQ,KAAK,CAAC,KAAK;AACnC,IAAM,WAAW,IAAI,cAAc;AAEnC,IAAI,YAAY,UAAU;AACxB,UAAQ,OAAO;AAAA,IACb,GAAG,KAAK,UAAU,EAAE,YAAY,kBAAkB,SAAS,MAAM,SAAS,OAAO,EAAE,GAAG,MAAM,CAAC,CAAC;AAAA;AAAA,EAChG;AACF,WAAW,YAAY,SAAS;AAC9B,UAAQ,OAAO;AAAA,IACb,GAAG,KAAK,UAAU,EAAE,YAAY,kBAAkB,SAAS,MAAM,eAAe,EAAE,GAAG,MAAM,CAAC,CAAC;AAAA;AAAA,EAC/F;AACF,WAAW,YAAY,WAAW;AAChC,QAAM,QAAQ,MAAM,SAAS;AAAA,IAC3B,EAAE,MAAM,OAAO,WAAW,GAAG,MAAM,sBAAsB;AAAA,IACzD,uBAAuB;AAAA,IACvB;AAAA,EACF;AACA,UAAQ,OAAO;AAAA,IACb,GAAG,KAAK,UAAU,EAAE,YAAY,kBAAkB,SAAS,MAAM,SAAS,QAAQ,KAAK,EAAE,GAAG,MAAM,CAAC,CAAC;AAAA;AAAA,EACtG;AACF,WAAW,YAAY,YAAY;AACjC,QAAM,QAAQ,MAAM,SAAS;AAAA,IAC3B,EAAE,MAAM,OAAO,aAAa,mBAAmB,WAAW,EAAE;AAAA,IAC5D,uBAAuB;AAAA,IACvB;AAAA,EACF;AACA,UAAQ,OAAO;AAAA,IACb,GAAG,KAAK,UAAU,EAAE,YAAY,kBAAkB,UAAU,MAAM,SAAS,SAAS,KAAK,EAAE,GAAG,MAAM,CAAC,CAAC;AAAA;AAAA,EACxG;AACF,OAAO;AACL,UAAQ,OAAO,MAAM,mCAAmC,OAAO;AAAA,CAAI;AACnE,UAAQ,WAAW;AACrB;","names":[]}

package/dist/index.d.ts

@@ -0,0 +1,45 @@
+import { GeneratedAsset, BrandContext, LicenseMetadata } from '@nebutra/generation-context';
+
+type AudioIntent = {
+    type: "bgm";
+    durationS: number;
+    mood: string;
+    bpm?: number;
+} | {
+    type: "sfx";
+    description: string;
+    durationS?: number;
+} | {
+    type: "song";
+    lyricsPrompt: string;
+    durationS: number;
+};
+interface AudioAsset extends GeneratedAsset {
+    readonly kind: "audio";
+    readonly durationS: number;
+    readonly format: "wav";
+    readonly loudnessLufs: number;
+}
+interface AudioHealth {
+    readonly provider: string;
+    readonly ok: boolean;
+    readonly suggestion?: string;
+}
+interface AudioPipelineOptions {
+    readonly root?: string;
+    readonly provider?: "tone-local" | "local-model" | "remote";
+}
+declare function readAudioDebug(root?: string, limit?: number): Promise<unknown[]>;
+declare function createToneWav(durationS: number, frequency?: number, sampleRate?: number): Buffer;
+declare class AudioPipeline {
+    #private;
+    constructor(options?: AudioPipelineOptions);
+    generate(intent: AudioIntent, brandInput: BrandContext | undefined, requireCommercial?: boolean): Promise<AudioAsset>;
+    doctor(): Promise<AudioHealth[]>;
+    license(asset: Pick<AudioAsset, "license">): Promise<LicenseMetadata>;
+    loudness(asset: Pick<AudioAsset, "loudnessLufs">): Promise<{
+        lufs: number;
+    }>;
+}
+
+export { type AudioAsset, type AudioHealth, type AudioIntent, AudioPipeline, type AudioPipelineOptions, createToneWav, readAudioDebug };

package/dist/index.js

@@ -0,0 +1,11 @@
+import {
+  AudioPipeline,
+  createToneWav,
+  readAudioDebug
+} from "./chunk-32LVUMTP.js";
+export {
+  AudioPipeline,
+  createToneWav,
+  readAudioDebug
+};
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":[],"sourcesContent":[],"mappings":"","names":[]}

package/examples/bgm.ts

@@ -0,0 +1,10 @@
+import { createDemoBrandContext } from "@nebutra/generation-context";
+import { AudioPipeline } from "../src/index";
+
+const asset = await new AudioPipeline().generate(
+  { type: "bgm", durationS: 5, mood: "uplifting tech", bpm: 110 },
+  createDemoBrandContext(),
+  true,
+);
+
+process.stdout.write(`${asset.path}\n`);

package/examples/license.ts

@@ -0,0 +1,11 @@
+import { createDemoBrandContext } from "@nebutra/generation-context";
+import { AudioPipeline } from "../src/index";
+
+const pipeline = new AudioPipeline();
+const asset = await pipeline.generate(
+  { type: "sfx", description: "keyboard typing", durationS: 2 },
+  createDemoBrandContext(),
+  true,
+);
+
+process.stdout.write(`${JSON.stringify(await pipeline.license(asset), null, 2)}\n`);

package/examples/loudness.ts

@@ -0,0 +1,11 @@
+import { createDemoBrandContext } from "@nebutra/generation-context";
+import { AudioPipeline } from "../src/index";
+
+const pipeline = new AudioPipeline();
+const asset = await pipeline.generate(
+  { type: "song", lyricsPrompt: "indie developer anthem", durationS: 8 },
+  createDemoBrandContext(),
+  true,
+);
+
+process.stdout.write(`${JSON.stringify(await pipeline.loudness(asset), null, 2)}\n`);

package/package.json

@@ -0,0 +1,59 @@
+{
+  "name": "@nebutra/audio-pipeline",
+  "version": "0.1.0",
+  "description": "BrandContext-first audio generation, license metadata, and loudness utilities",
+  "private": false,
+  "license": "MIT",
+  "type": "module",
+  "nebutra": {
+    "status": "wip",
+    "productionReady": false,
+    "surface": "generation-capability",
+    "requires": [
+      "@nebutra/generation-context for BrandContext",
+      "@nebutra/sandbox-runtime for model and ffmpeg sidecars"
+    ],
+    "gaps": [
+      "Music model adapters are health-checked but not launched by this package",
+      "Commercial license policy is local metadata until asset governance lands",
+      "LUFS measurement uses deterministic estimation until ffmpeg/pyloudnorm sidecar is wired"
+    ],
+    "featureId": "audio-pipeline",
+    "category": "ai",
+    "summary": "License-aware BGM, SFX, and song generation surface"
+  },
+  "main": "./src/index.ts",
+  "types": "./src/index.ts",
+  "exports": {
+    ".": "./src/index.ts"
+  },
+  "dependencies": {
+    "@nebutra/capability-kit": "0.2.0",
+    "@nebutra/errors": "0.1.0",
+    "@nebutra/generation-context": "0.1.0"
+  },
+  "devDependencies": {
+    "@types/node": "^22.19.15",
+    "tsup": "^8.5.1",
+    "tsx": "^4.21.0",
+    "typescript": "^5.9.3",
+    "vitest": "^4.1.4"
+  },
+  "homepage": "https://github.com/Nebutra/Nebutra-Sailor/tree/main/packages/ai/audio-pipeline#readme",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/Nebutra/Nebutra-Sailor.git",
+    "directory": "packages/ai/audio-pipeline"
+  },
+  "bugs": {
+    "url": "https://github.com/Nebutra/Nebutra-Sailor/issues"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "scripts": {
+    "build": "tsup",
+    "test": "vitest run",
+    "typecheck": "tsc --noEmit"
+  }
+}

package/src/cli.ts

@@ -0,0 +1,36 @@
+import { createDemoBrandContext } from "@nebutra/generation-context";
+import { AudioPipeline, readAudioDebug } from "./index";
+
+const command = process.argv[2] ?? "doctor";
+const pipeline = new AudioPipeline();
+
+if (command === "doctor") {
+  process.stdout.write(
+    `${JSON.stringify({ capability: "audio-pipeline", results: await pipeline.doctor() }, null, 2)}\n`,
+  );
+} else if (command === "debug") {
+  process.stdout.write(
+    `${JSON.stringify({ capability: "audio-pipeline", entries: await readAudioDebug() }, null, 2)}\n`,
+  );
+} else if (command === "license") {
+  const asset = await pipeline.generate(
+    { type: "bgm", durationS: 3, mood: "uplifting technical" },
+    createDemoBrandContext(),
+    true,
+  );
+  process.stdout.write(
+    `${JSON.stringify({ capability: "audio-pipeline", license: await pipeline.license(asset) }, null, 2)}\n`,
+  );
+} else if (command === "loudness") {
+  const asset = await pipeline.generate(
+    { type: "sfx", description: "keyboard typing", durationS: 1 },
+    createDemoBrandContext(),
+    true,
+  );
+  process.stdout.write(
+    `${JSON.stringify({ capability: "audio-pipeline", loudness: await pipeline.loudness(asset) }, null, 2)}\n`,
+  );
+} else {
+  process.stderr.write(`Unknown audio-pipeline command: ${command}\n`);
+  process.exitCode = 1;
+}

package/src/index.test.ts

@@ -0,0 +1,33 @@
+import { mkdtemp, readFile, rm } from "node:fs/promises";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+import { createDemoBrandContext } from "@nebutra/generation-context";
+import { afterEach, describe, expect, it } from "vitest";
+import { AudioPipeline, readAudioDebug } from "./index";
+
+let root: string | undefined;
+
+afterEach(async () => {
+  if (root) await rm(root, { recursive: true, force: true });
+  root = undefined;
+});
+
+describe("AudioPipeline", () => {
+  it("requires BrandContext", async () => {
+    await expect(
+      new AudioPipeline().generate({ type: "sfx", description: "click" }, undefined, true),
+    ).rejects.toMatchObject({ capability: "audio-pipeline" });
+  });
+
+  it("writes a valid WAV header with commercial metadata", async () => {
+    root = await mkdtemp(join(tmpdir(), "audio-pipeline-"));
+    const asset = await new AudioPipeline({ root }).generate(
+      { type: "bgm", durationS: 4, mood: "calm technical" },
+      createDemoBrandContext(),
+      true,
+    );
+    expect((await readFile(asset.path)).subarray(0, 4).toString()).toBe("RIFF");
+    expect(asset.license.status).toBe("commercial-ok");
+    expect(await readAudioDebug(root)).toHaveLength(1);
+  });
+});

package/src/index.ts

@@ -0,0 +1,176 @@
+import { mkdir, writeFile } from "node:fs/promises";
+import { dirname, join } from "node:path";
+import { appendCapabilityDebug, readCapabilityDebug } from "@nebutra/capability-kit/debug";
+import { CapabilityError } from "@nebutra/errors";
+import {
+  assetId,
+  type BrandContext,
+  type GeneratedAsset,
+  type LicenseMetadata,
+  requireBrandContext,
+} from "@nebutra/generation-context";
+
+export type AudioIntent =
+  | { type: "bgm"; durationS: number; mood: string; bpm?: number }
+  | { type: "sfx"; description: string; durationS?: number }
+  | { type: "song"; lyricsPrompt: string; durationS: number };
+
+export interface AudioAsset extends GeneratedAsset {
+  readonly kind: "audio";
+  readonly durationS: number;
+  readonly format: "wav";
+  readonly loudnessLufs: number;
+}
+
+export interface AudioHealth {
+  readonly provider: string;
+  readonly ok: boolean;
+  readonly suggestion?: string;
+}
+
+export interface AudioPipelineOptions {
+  readonly root?: string;
+  readonly provider?: "tone-local" | "local-model" | "remote";
+}
+
+export async function readAudioDebug(root = process.cwd(), limit = 10): Promise<unknown[]> {
+  return readCapabilityDebug("audio-pipeline", { root, limit });
+}
+
+function intentLabel(intent: AudioIntent): string {
+  switch (intent.type) {
+    case "bgm":
+      return intent.mood;
+    case "sfx":
+      return intent.description;
+    case "song":
+      return intent.lyricsPrompt;
+  }
+}
+
+function durationFor(intent: AudioIntent): number {
+  if (intent.type === "sfx") return Math.max(1, Math.min(intent.durationS ?? 3, 8));
+  return Math.max(1, Math.min(intent.durationS, 12));
+}
+
+export function createToneWav(durationS: number, frequency = 440, sampleRate = 24_000): Buffer {
+  const samples = Math.max(1, Math.floor(durationS * sampleRate));
+  const dataBytes = samples * 2;
+  const buffer = Buffer.alloc(44 + dataBytes);
+  buffer.write("RIFF", 0);
+  buffer.writeUInt32LE(36 + dataBytes, 4);
+  buffer.write("WAVE", 8);
+  buffer.write("fmt ", 12);
+  buffer.writeUInt32LE(16, 16);
+  buffer.writeUInt16LE(1, 20);
+  buffer.writeUInt16LE(1, 22);
+  buffer.writeUInt32LE(sampleRate, 24);
+  buffer.writeUInt32LE(sampleRate * 2, 28);
+  buffer.writeUInt16LE(2, 32);
+  buffer.writeUInt16LE(16, 34);
+  buffer.write("data", 36);
+  buffer.writeUInt32LE(dataBytes, 40);
+  for (let index = 0; index < samples; index += 1) {
+    const fade = Math.min(index / 1200, (samples - index) / 1200, 1);
+    const sample = Math.round(
+      Math.sin((2 * Math.PI * frequency * index) / sampleRate) * 9000 * fade,
+    );
+    buffer.writeInt16LE(sample, 44 + index * 2);
+  }
+  return buffer;
+}
+
+function frequencyFor(brand: BrandContext, intent: AudioIntent): number {
+  const base = brand.toneKeywords.join("").length + intentLabel(intent).length;
+  return 220 + (base % 9) * 55;
+}
+
+function licenseFor(requireCommercial: boolean, provider: string): LicenseMetadata {
+  if (requireCommercial && provider !== "tone-local") {
+    return {
+      status: "unknown",
+      source: provider,
+      suggestion: "Verify the provider plan and model terms before commercial use.",
+    };
+  }
+  return { status: "commercial-ok", source: "deterministic local renderer" };
+}
+
+export class AudioPipeline {
+  readonly #root: string;
+  readonly #provider: "tone-local" | "local-model" | "remote";
+
+  constructor(options: AudioPipelineOptions = {}) {
+    this.#root = options.root ?? process.cwd();
+    this.#provider = options.provider ?? "tone-local";
+  }
+
+  async generate(
+    intent: AudioIntent,
+    brandInput: BrandContext | undefined,
+    requireCommercial = true,
+  ): Promise<AudioAsset> {
+    const brand = requireBrandContext(brandInput, "audio-pipeline");
+    if (requireCommercial && this.#provider !== "tone-local") {
+      throw new CapabilityError(
+        "audio-pipeline",
+        "Selected audio provider lacks verified license",
+        {
+          suggestion: "Use tone-local or wire a provider with commercial license metadata.",
+          statusCode: 409,
+        },
+      );
+    }
+    const durationS = durationFor(intent);
+    const id = assetId("audio", `${intent.type}_${intentLabel(intent)}`);
+    const path = join(this.#root, ".nebutra", "generated", "audio-pipeline", `${id}.wav`);
+    await mkdir(dirname(path), { recursive: true });
+    await writeFile(path, createToneWav(durationS, frequencyFor(brand, intent)));
+
+    const asset: AudioAsset = {
+      id,
+      tenantId: brand.tenantId,
+      kind: "audio",
+      path,
+      brandId: brand.brandId,
+      provider: this.#provider,
+      model: "brand-context-tone-v1",
+      createdAt: new Date().toISOString(),
+      license: licenseFor(requireCommercial, this.#provider),
+      durationS,
+      format: "wav",
+      loudnessLufs: -18,
+      metadata: { intent, brandSource: brand.sourcePath },
+    };
+    await appendCapabilityDebug(
+      "audio-pipeline",
+      { type: "generate", asset },
+      { root: this.#root },
+    );
+    return asset;
+  }
+
+  async doctor(): Promise<AudioHealth[]> {
+    return [
+      { provider: "tone-local", ok: true },
+      {
+        provider: "local-model",
+        ok: Boolean(process.env.AUDIO_LOCAL_MODEL_PATH),
+        suggestion: "Set AUDIO_LOCAL_MODEL_PATH to enable local model generation.",
+      },
+      {
+        provider: "remote",
+        ok: Boolean(process.env.AUDIO_REMOTE_API_KEY),
+        suggestion: "Set AUDIO_REMOTE_API_KEY to enable remote audio fallback.",
+      },
+    ];
+  }
+
+  async license(asset: Pick<AudioAsset, "license">): Promise<LicenseMetadata> {
+    return asset.license;
+  }
+
+  async loudness(asset: Pick<AudioAsset, "loudnessLufs">): Promise<{ lufs: number }> {
+    return { lufs: asset.loudnessLufs };
+  }
+}

package/tsconfig.json

@@ -0,0 +1,12 @@
+{
+  "extends": "../../../tsconfig.base.json",
+  "compilerOptions": {
+    "module": "ESNext",
+    "moduleResolution": "bundler",
+    "target": "esnext",
+    "types": ["node"],
+    "incremental": false
+  },
+  "include": ["src", "examples"],
+  "exclude": ["node_modules", "dist"]
+}

package/tsup.config.ts

@@ -0,0 +1,11 @@
+import { defineConfig } from "tsup";
+
+export default defineConfig({
+  entry: ["src/index.ts", "src/cli.ts"],
+  format: ["esm"],
+  dts: true,
+  sourcemap: true,
+  clean: true,
+  target: "es2022",
+  external: ["@nebutra/capability-kit/debug", "@nebutra/errors", "@nebutra/generation-context"],
+});