@doufunao123/asset-gateway 0.11.0 → 0.12.0

package/dist/index.js

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
 
 // src/index.ts
-import { Command as Command9 } from "commander";
+import { Command as Command10 } from "commander";
 
 // src/commands/auth.ts
 import { existsSync as existsSync2, unlinkSync } from "fs";
@@ -70,7 +70,7 @@ function normalizeError(error2) {
 
 // src/meta.ts
 var CLI_NAME = "asset-gateway";
-var CLI_VERSION = "0.8.1";
+var CLI_VERSION = "0.11.1";
 var CLI_DESCRIPTION = "Universal asset generation gateway CLI";
 var DEFAULT_GATEWAY_URL = "https://upload.xiaomao.chat";
 
@@ -174,6 +174,8 @@ var GatewayClient = class {
     this.baseUrl = baseUrl;
     this.token = token;
   }
+  baseUrl;
+  token;
   async get(path) {
     return this.request("GET", path);
   }
@@ -356,13 +358,17 @@ function mask(token) {
 
 // src/commands/describe.ts
 import { Command as Command2 } from "commander";
+
+// src/describe-schemas.ts
 var SCHEMAS = {
   auth: {
     description: "Credential management",
     subcommands: {
       set: {
         description: "Save token and gateway URL locally",
-        params: { token: { type: "string", required: true, description: "API token (agk_ prefix for admin, or any API key)" } }
+        params: {
+          token: { type: "string", required: true, description: "API token or admin token" }
+        }
       },
       status: { description: "Show current authentication status" },
       clear: { description: "Remove saved credentials" }
@@ -372,91 +378,222 @@ var SCHEMAS = {
     description: "Generate assets via the gateway",
     subcommands: {
       image: {
-        description: "Generate an image from a text prompt",
+        description: "Generate or edit an image (Gemini / GPT Image / Grok)",
         params: {
-          "--prompt": { type: "string", required: true, description: "Image description prompt" },
-          "--provider": { type: "string", required: false, description: "Provider to use" },
-          "--transparent": { type: "bool", default: false, description: "Request transparent background" },
-          "--model": { type: "string", required: false, description: "Model to use" },
-          "--size": { type: "string", required: false, description: "Image size (e.g. 1024x1024)" },
-          "--output-dir": { type: "string", default: ".", description: "Directory to save output" }
+          "--prompt": { type: "string", required: true, description: "Image prompt" },
+          "--provider": { type: "string", required: false },
+          "--transparent": { type: "bool", description: "Transparent background" },
+          "--model": { type: "string" },
+          "--size": { type: "string", description: 'e.g. "1024x1024"' },
+          "--input": { type: "string", description: "Image URL for editing" },
+          "--ref": { type: "string[]", description: "Reference image URLs (repeatable)" },
+          "--edit-mode": { type: "string", description: "edit | inpaint | restyle | expand" },
+          "--session": { type: "string", description: "Multi-turn session id" },
+          "--output-dir": { type: "string", default: "." }
         }
       },
       video: {
-        description: "Generate a video from a text prompt",
+        description: "Text or image-to-video (Grok)",
+        params: {
+          "--prompt": { type: "string", required: true },
+          "--provider": { type: "string" },
+          "--input": { type: "string", description: "Image URL for I2V" },
+          "--output-dir": { type: "string", default: "." }
+        }
+      },
+      batch: {
+        description: "Batch image (etc.) generation; optional sprite compose",
         params: {
-          "--prompt": { type: "string", required: true, description: "Video description prompt" },
-          "--provider": { type: "string", required: false, description: "Provider to use" },
-          "--output-dir": { type: "string", default: ".", description: "Directory to save output" }
+          "--prompt": { type: "string[]", required: true, description: "One prompt per frame" },
+          "--asset-type": { type: "string", default: "image" },
+          "--transparent": { type: "bool" },
+          "--size": { type: "string" },
+          "--ref": { type: "string[]" },
+          "--compose": { type: "string", description: "horizontal | vertical | grid" },
+          "--columns": { type: "number" },
+          "--frame-size": { type: "string", description: "e.g. 64x64" },
+          "--output-dir": { type: "string", default: "." }
         }
       },
       audio: {
-        description: "Generate audio from a text prompt",
+        description: "BGM/SFX via ElevenLabs sound-generation",
+        params: {
+          "--prompt": { type: "string", required: true },
+          "--type": { type: "string", description: "bgm | sfx" },
+          "--duration": { type: "number", description: "Seconds" },
+          "--output-dir": { type: "string", default: "." }
+        }
+      },
+      music: {
+        description: "Music via ElevenLabs POST /v1/music (model music_v1)",
+        params: {
+          "--prompt": { type: "string", required: true },
+          "--duration": { type: "number", description: "Seconds; mapped to music_length_ms (3s\u2013600s)" },
+          "--force-instrumental": { type: "bool", description: "Optional; passed to provider" },
+          "--output-format": { type: "string", description: "Optional; e.g. mp3_44100_128" },
+          "--output-dir": { type: "string", default: "." }
+        }
+      },
+      tts: {
+        description: "TTS: default Qwen3-TTS (voice/language/instructions). ElevenLabs: --provider elevenlabs --voice-id <id>",
         params: {
-          "--prompt": { type: "string", required: true, description: "Audio description prompt" },
-          "--type": { type: "string", required: false, description: "Audio type: bgm or sfx" },
-          "--duration": { type: "number", required: false, description: "Duration in seconds" },
-          "--output-dir": { type: "string", default: ".", description: "Directory to save output" }
+          "--prompt": { type: "string", required: true },
+          "--voice": { type: "string", description: "Qwen voice name or custom id", default: "Cherry" },
+          "--voice-id": { type: "string", description: "ElevenLabs voice id (with --provider elevenlabs)" },
+          "--language": { type: "string", default: "Auto" },
+          "--model": { type: "string", default: "qwen3-tts-flash" },
+          "--instructions": { type: "string", description: "Instruct-model style control" },
+          "--provider": { type: "string", description: "qwen_tts | elevenlabs" },
+          "--output-dir": { type: "string", default: "." }
         }
       },
       model: {
-        description: "Generate a 3D model",
+        description: "3D model generation (Tripo)",
         params: {
-          "--image": { type: "string", required: false, description: "Reference image URL" },
-          "--prompt": { type: "string", required: false, description: "Model description prompt" },
-          "--output-dir": { type: "string", default: ".", description: "Directory to save output" }
+          "--image": { type: "string", description: "Reference image URL" },
+          "--prompt": { type: "string" },
+          "--model-version": { type: "string" },
+          "--face-limit": { type: "number" },
+          "--pbr": { type: "bool" },
+          "--texture-quality": { type: "string" },
+          "--auto-size": { type: "bool" },
+          "--negative-prompt": { type: "string" },
+          "--multiview": { type: "string", description: "Comma-separated 4 view URLs" },
+          "--output-dir": { type: "string", default: "." }
         }
       },
       text: {
-        description: "Generate text via LLM",
+        description: "LLM text via proxy",
         params: {
-          "--prompt": { type: "string", required: true, description: "Text prompt" },
-          "--model": { type: "string", required: false, description: "Model to use" },
-          "--max-tokens": { type: "number", required: false, description: "Maximum tokens" },
-          "--output-dir": { type: "string", default: ".", description: "Directory to save output" }
+          "--prompt": { type: "string", required: true },
+          "--model": { type: "string" },
+          "--max-tokens": { type: "number" },
+          "--output-dir": { type: "string", default: "." }
         }
       }
     }
   },
-  provider: {
-    description: "Provider management",
+  process: {
+    description: "Image/video post-process on gateway (ImageMagick/ffmpeg/rembg)",
     subcommands: {
-      list: { description: "List available providers" },
-      health: {
-        description: "Check provider health",
-        params: { name: { type: "string", required: false, description: "Specific provider name" } }
+      crop: {
+        description: "Smart crop (trim / power_of2)",
+        params: {
+          "--input": { type: "string", required: true, description: "File path or URL" },
+          "--mode": { type: "string", default: "tightest" },
+          "--output-dir": { type: "string", default: "." }
+        }
+      },
+      resize: {
+        description: "Resize to exact width/height",
+        params: {
+          "--input": { type: "string", required: true },
+          "--width": { type: "number", required: true },
+          "--height": { type: "number", required: true },
+          "--output-dir": { type: "string", default: "." }
+        }
+      },
+      compose: {
+        description: "Sprite sheet from multiple images",
+        params: {
+          "--input": { type: "string[]", required: true },
+          "--direction": { type: "string", default: "horizontal" },
+          "--columns": { type: "number" },
+          "--padding": { type: "string" },
+          "--frame-width": { type: "number" },
+          "--frame-height": { type: "number" },
+          "--output-dir": { type: "string", default: "." }
+        }
+      },
+      "extract-frames": {
+        description: "Sample frames from video (ffmpeg)",
+        params: {
+          "--input": { type: "string", required: true },
+          "--count": { type: "string", default: "8" },
+          "--output-dir": { type: "string", default: "." }
+        }
+      },
+      "remove-bg": {
+        description: "Remove background (rembg / fallback)",
+        params: {
+          "--input": { type: "string[]", required: true },
+          "--bg-color": { type: "string" },
+          "--output-dir": { type: "string", default: "." }
+        }
       }
     }
   },
+  process3d: {
+    description: "Tripo 3D follow-up operations (chain on tripo_task_id)",
+    subcommands: {
+      convert: { description: "Export format (FBX/GLTF/\u2026)", params: { "--task-id": { required: true }, "--format": { required: true } } },
+      texture: { description: "Re-texture", params: { "--task-id": { required: true } } },
+      rig: { description: "Auto-rig", params: { "--task-id": { required: true } } },
+      animate: { description: "Retarget animation", params: { "--task-id": { required: true } } },
+      "render-sprites": { description: "Blender render to 2D frames", params: { "--task-id": { required: true } } },
+      reduce: { description: "Low-poly", params: { "--task-id": { required: true } } },
+      stylize: { description: "Style transfer", params: { "--task-id": { required: true } } },
+      segment: { description: "Mesh segmentation", params: { "--task-id": { required: true } } },
+      prerigcheck: { description: "Rig eligibility", params: { "--task-id": { required: true } } },
+      refine: { description: "Refine quality", params: { "--task-id": { required: true } } },
+      import: { description: "Import external model", params: { "--file-url": { type: "string" }, "--file-path": { type: "string" } } }
+    }
+  },
+  voice: {
+    description: "Qwen3-TTS custom voices (clone / design / list / delete)",
+    subcommands: {
+      clone: {
+        description: "Clone from audio sample",
+        params: { "--audio": { required: true }, "--name": { required: true } }
+      },
+      design: {
+        description: "Design voice from text",
+        params: { "--prompt": { required: true }, "--name": { required: true } }
+      },
+      list: { description: "List custom voices", params: { "--type": { type: "string" } } },
+      delete: { description: "Delete by voice id", params: { "<voice-id>": { required: true } } }
+    }
+  },
+  upload: {
+    description: "Upload and list gateway assets",
+    subcommands: {
+      file: { description: "Upload file \u2192 URL", params: { "<path>": { required: true } } },
+      list: { description: "List uploads" },
+      delete: { description: "Delete by filename (admin)", params: { "<filename>": { required: true } } }
+    }
+  },
+  provider: {
+    description: "Provider discovery and health",
+    subcommands: {
+      list: { description: "List providers" },
+      health: { description: "Health check", params: { name: { type: "string", required: false } } }
+    }
+  },
   job: {
-    description: "Job management",
+    description: "Async job history",
     subcommands: {
       list: {
         description: "List jobs",
         params: {
-          "--status": { type: "string", required: false, description: "Filter by status" },
-          "--limit": { type: "number", required: false, description: "Maximum number of jobs to return" }
+          "--status": { type: "string" },
+          "--limit": { type: "number" }
         }
       },
-      status: {
-        description: "Get job status",
-        params: { id: { type: "string", required: true, description: "Job ID" } }
-      },
-      cancel: {
-        description: "Cancel a job",
-        params: { id: { type: "string", required: true, description: "Job ID" } }
-      }
+      status: { description: "Job detail", params: { id: { type: "string", required: true } } },
+      cancel: { description: "Cancel pending/running", params: { id: { type: "string", required: true } } }
     }
   },
   describe: {
-    description: "Self-describe available commands (JSON Schema)",
+    description: "Command introspection (this output)",
     params: {
-      command: { type: "string", required: false, description: "Specific command to describe" }
+      command: { type: "string", required: false, description: "Top-level group: generate, process, \u2026" }
     }
   }
 };
+
+// src/commands/describe.ts
 function createDescribeCommand() {
-  return new Command2("describe").description("Self-describe available commands (JSON Schema)").argument("[command]", "Specific command to describe").action(function(commandArg) {
+  return new Command2("describe").description("Self-describe available commands (JSON Schema)").argument("[command]", "Specific command group to describe (e.g. generate, process)").action(function(commandArg) {
     const globals = this.optsWithGlobals();
     if (!commandArg) {
       output(
@@ -487,10 +624,10 @@ function createDescribeCommand() {
 
 // src/commands/generate.ts
 import { mkdirSync as mkdirSync2, writeFileSync as writeFileSync2 } from "fs";
-import { join as join2 } from "path";
+import { dirname as dirname2, join as join2 } from "path";
 import { Command as Command3 } from "commander";
 function inferExtension(assetType) {
-  const map = { image: "png", audio: "mp3", music: "mp3", tts: "mp3", video: "mp4", model3d: "glb", text: "txt" };
+  const map = { image: "png", audio: "mp3", music: "mp3", tts: "mp3", video: "mp4", model3d: "glb", text: "txt", sprite: "png" };
   return map[assetType] ?? "bin";
 }
 function stripDataUri(data) {
@@ -524,6 +661,37 @@ async function saveOutput(result, assetType, outputDir) {
   }
   return null;
 }
+async function saveNamedOutput(result, assetType, filePath) {
+  mkdirSync2(dirname2(filePath), { recursive: true });
+  if (result.output_data) {
+    const raw = String(result.output_data);
+    if (assetType === "text") {
+      writeFileSync2(filePath, raw, "utf8");
+    } else {
+      writeFileSync2(filePath, Buffer.from(stripDataUri(raw), "base64"));
+    }
+    return filePath;
+  }
+  if (result.output_url) {
+    const response = await fetch(String(result.output_url));
+    if (!response.ok) {
+      return null;
+    }
+    const buffer = Buffer.from(await response.arrayBuffer());
+    writeFileSync2(filePath, buffer);
+    return filePath;
+  }
+  return null;
+}
+function parseFrameSize(raw) {
+  const [width, height] = raw.split("x");
+  const frameWidth = Number(width);
+  const frameHeight = Number(height);
+  if (!Number.isFinite(frameWidth) || !Number.isFinite(frameHeight)) {
+    throw new Error("frame size must be formatted as WIDTHxHEIGHT");
+  }
+  return { frame_width: frameWidth, frame_height: frameHeight };
+}
 function createGenerateCommand() {
   const command = new Command3("generate").description("Generate assets via the gateway");
   command.addCommand(
@@ -552,7 +720,7 @@ function createGenerateCommand() {
     })
   );
   command.addCommand(
-    new Command3("video").description("Generate a video from a text prompt").requiredOption("--prompt <text>", "Video description prompt").option("--provider <id>", "Provider to use").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
+    new Command3("video").description("Generate a video from a text prompt (or image-to-video with --input)").requiredOption("--prompt <text>", "Video description prompt").option("--provider <id>", "Provider to use").option("--input <url>", "Reference image URL for image-to-video (Grok)").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
       try {
         const ctx = createContext(this);
         const body = {
@@ -560,6 +728,7 @@ function createGenerateCommand() {
           prompt: options.prompt
         };
         if (options.provider) body.provider = options.provider;
+        if (options.input) body.input_file = options.input;
         const data = await ctx.client.post("/api/generate", body);
         const localPath = await saveOutput(data, "video", options.outputDir);
         if (localPath) data.local_path = localPath;
@@ -569,6 +738,56 @@ function createGenerateCommand() {
       }
     })
   );
+  command.addCommand(
+    new Command3("batch").description("Batch generate multiple assets with shared parameters").requiredOption("--prompt <texts...>", "Multiple prompts (one per frame)").option("--asset-type <type>", "Asset type", "image").option("--transparent", "Request transparent background").option("--size <size>", "Image size").option("--ref <urls...>", "Reference image URLs").option("--compose <direction>", "Auto-compose: horizontal, vertical, grid").option("--columns <n>", "Grid columns for compose").option("--frame-size <size>", "Frame size for compose (e.g. 64x64)").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const shared = {};
+        if (options.transparent) shared.transparent = true;
+        if (options.size) shared.size = options.size;
+        if (options.ref && options.ref.length > 0) shared.reference_images = options.ref;
+        const body = {
+          asset_type: options.assetType,
+          prompts: options.prompt,
+          shared
+        };
+        if (options.compose) {
+          const compose = { direction: options.compose };
+          if (options.columns) compose.columns = Number(options.columns);
+          if (options.frameSize) Object.assign(compose, parseFrameSize(options.frameSize));
+          body.compose = compose;
+        }
+        const data = await ctx.client.post("/api/generate/batch", body);
+        const assetType = String(options.assetType);
+        mkdirSync2(options.outputDir, { recursive: true });
+        if (Array.isArray(data.frames)) {
+          for (const frame of data.frames) {
+            if (typeof frame !== "object" || frame === null) continue;
+            const record = frame;
+            const index = typeof record.index === "number" ? record.index : 0;
+            const localPath = await saveNamedOutput(
+              record,
+              assetType,
+              join2(options.outputDir, `frame_${String(index).padStart(3, "0")}.${inferExtension(assetType)}`)
+            );
+            if (localPath) record.local_path = localPath;
+          }
+        }
+        if (typeof data.spritesheet === "object" && data.spritesheet !== null) {
+          const record = data.spritesheet;
+          const localPath = await saveNamedOutput(
+            record,
+            "image",
+            join2(options.outputDir, "spritesheet.png")
+          );
+          if (localPath) record.local_path = localPath;
+        }
+        printSuccess("generate.batch", data, ctx);
+      } catch (error2) {
+        printError("generate.batch", error2);
+      }
+    })
+  );
   command.addCommand(
     new Command3("audio").description("Generate audio from a text prompt").requiredOption("--prompt <text>", "Audio description prompt").option("--type <type>", "Audio type: bgm or sfx").option("--duration <seconds>", "Duration in seconds").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
       try {
@@ -591,16 +810,21 @@ function createGenerateCommand() {
     })
   );
   command.addCommand(
-    new Command3("music").description("Generate music from a text prompt").requiredOption("--prompt <text>", "Music description prompt").option("--duration <seconds>", "Duration in seconds").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
+    new Command3("music").description("Generate music (ElevenLabs /v1/music)").requiredOption("--prompt <text>", "Music description prompt").option("--duration <seconds>", "Duration in seconds (maps to music_length_ms)").option("--force-instrumental", "Force instrumental output (ElevenLabs)").option(
+      "--output-format <fmt>",
+      "ElevenLabs output_format query, e.g. mp3_44100_128"
+    ).option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
       try {
         const ctx = createContext(this);
         const body = {
           asset_type: "music",
           prompt: options.prompt
         };
-        if (options.duration) {
-          body.params = { duration_seconds: Number(options.duration) };
-        }
+        const params = {};
+        if (options.duration) params.duration_seconds = Number(options.duration);
+        if (options.forceInstrumental) params.force_instrumental = true;
+        if (options.outputFormat) params.output_format = options.outputFormat;
+        if (Object.keys(params).length > 0) body.params = params;
         const data = await ctx.client.post("/api/generate", body);
         const localPath = await saveOutput(data, "music", options.outputDir);
         if (localPath) data.local_path = localPath;
@@ -611,7 +835,12 @@ function createGenerateCommand() {
     })
   );
   command.addCommand(
-    new Command3("tts").description("Text-to-speech synthesis via Qwen3-TTS").requiredOption("--prompt <text>", "Text to synthesize").option("--voice <name>", "Voice name or custom voice ID", "Cherry").option("--language <lang>", "Language hint: Auto, Chinese, English, Japanese, etc.", "Auto").option("--model <model>", "Qwen3-TTS model", "qwen3-tts-flash").option("--instructions <text>", "Natural language speaking instructions (for instruct models)").option("--provider <id>", "Provider to use").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
+    new Command3("tts").description(
+      "Text-to-speech: default Qwen3-TTS; use --provider elevenlabs --voice-id for ElevenLabs"
+    ).requiredOption("--prompt <text>", "Text to synthesize").option("--voice <name>", "Qwen voice name or custom voice id", "Cherry").option(
+      "--voice-id <id>",
+      "ElevenLabs voice_id (use with --provider elevenlabs; routes to TTS API)"
+    ).option("--language <lang>", "Language hint: Auto, Chinese, English, Japanese, etc.", "Auto").option("--model <model>", "Model id (Qwen TTS or ElevenLabs model_id)", "qwen3-tts-flash").option("--instructions <text>", "Natural language speaking instructions (for instruct models)").option("--provider <id>", "qwen_tts | elevenlabs").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
       try {
         const ctx = createContext(this);
         const params = {
@@ -619,6 +848,7 @@ function createGenerateCommand() {
           language_type: options.language
         };
         if (options.instructions) params.instructions = options.instructions;
+        if (options.voiceId) params.voice_id = options.voiceId;
         const body = {
           asset_type: "tts",
           prompt: options.prompt,
@@ -683,6 +913,34 @@ function createGenerateCommand() {
       }
     })
   );
+  command.addCommand(
+    new Command3("sprite").description("Animate a sprite image using PixelEngine AI").requiredOption("--prompt <text>", "Animation prompt describing the desired motion").requiredOption("--input <path>", "Input sprite image (local path or URL)").option("--model <model>", "Model: pixel-engine-v1.1 or frame-engine-v1.1", "pixel-engine-v1.1").option("--output-frames <n>", "Number of animation frames (even integer)", "8").option("--output-format <fmt>", "Output format: spritesheet, webp, gif", "spritesheet").option("--colors <n>", "Pixel palette color count (2-256, pixel model only)").option("--negative-prompt <text>", "What to avoid in the generation").option("--seed <n>", "Seed for reproducibility").option("--matte-color <hex>", "Matte color for alpha flattening (6-char hex)").option("--enhance-prompt", "Enhance the prompt before generation").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const params = {
+          output_frames: Number(options.outputFrames),
+          output_format: options.outputFormat
+        };
+        if (options.colors) params.pixel_config = { colors: Number(options.colors) };
+        if (options.negativePrompt) params.negative_prompt = options.negativePrompt;
+        if (options.seed) params.seed = Number(options.seed);
+        if (options.matteColor) params.matte_color = options.matteColor;
+        const body = {
+          asset_type: "sprite",
+          prompt: options.prompt,
+          model: options.model,
+          input_file: options.input,
+          params
+        };
+        const data = await ctx.client.post("/api/generate", body);
+        const localPath = await saveOutput(data, "sprite", options.outputDir);
+        if (localPath) data.local_path = localPath;
+        printSuccess("generate.sprite", data, ctx);
+      } catch (error2) {
+        printError("generate.sprite", error2);
+      }
+    })
+  );
   return command;
 }
 
@@ -744,17 +1002,32 @@ function readInputAsBase64(input) {
   return input;
 }
 function saveProcessOutput(data, outputDir) {
-  const outputData = data.output_data;
-  if (typeof outputData !== "string" || !outputData) return null;
   mkdirSync3(outputDir, { recursive: true });
   const timestamp = Date.now();
+  const outputs = data.outputs;
+  if (Array.isArray(outputs) && outputs.length > 0) {
+    const localPaths = [];
+    for (let i = 0; i < outputs.length; i++) {
+      const item = outputs[i];
+      const b64 = item.output_data;
+      if (typeof b64 !== "string" || !b64) continue;
+      const filePath2 = join3(outputDir, `frame_${timestamp}_${String(i).padStart(4, "0")}.png`);
+      writeFileSync3(filePath2, Buffer.from(b64, "base64"));
+      localPaths.push(filePath2);
+    }
+    delete data.outputs;
+    data.local_paths = localPaths;
+    return localPaths[0] ?? null;
+  }
+  const outputData = data.output_data;
+  if (typeof outputData !== "string" || !outputData) return null;
   const filePath = join3(outputDir, `processed_${timestamp}.png`);
   writeFileSync3(filePath, Buffer.from(outputData, "base64"));
   delete data.output_data;
   return filePath;
 }
 function createProcessCommand() {
-  const command = new Command5("process").description("Post-process images (crop, resize)");
+  const command = new Command5("process").description("Post-process images and video (crop, resize, compose, extract-frames, remove-bg)");
   command.addCommand(
     new Command5("crop").description("Smart crop an image (trim transparent borders)").requiredOption("--input <path>", "Input image (file path or URL)").option("--mode <mode>", "Crop mode: tightest or power_of2", "tightest").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
       try {
@@ -787,6 +1060,65 @@ function createProcessCommand() {
       }
     })
   );
+  command.addCommand(
+    new Command5("compose").description("Compose multiple images into a sprite sheet").requiredOption("--input <paths...>", "Input images (files or URLs)").option("--direction <dir>", "Layout: horizontal, vertical, grid", "horizontal").option("--columns <n>", "Columns for grid layout").option("--padding <n>", "Padding between frames in px", "0").option("--frame-width <n>", "Normalize each frame to this width").option("--frame-height <n>", "Normalize each frame to this height").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const inputs = Array.isArray(options.input) ? options.input : [options.input];
+        const data = await ctx.client.post("/api/process", {
+          inputs: inputs.map(readInputAsBase64),
+          operations: [{
+            op: "compose",
+            direction: options.direction,
+            columns: options.columns ? Number(options.columns) : void 0,
+            padding: Number(options.padding),
+            frame_width: options.frameWidth ? Number(options.frameWidth) : void 0,
+            frame_height: options.frameHeight ? Number(options.frameHeight) : void 0
+          }]
+        });
+        const localPath = saveProcessOutput(data, options.outputDir);
+        if (localPath) data.local_path = localPath;
+        printSuccess("process.compose", data, ctx);
+      } catch (error2) {
+        printError("process.compose", error2);
+      }
+    })
+  );
+  command.addCommand(
+    new Command5("extract-frames").description("Extract evenly-spaced frames from a video using ffmpeg").requiredOption("--input <path>", "Input video (file path or URL)").option("--count <n>", "Number of frames to extract", "8").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const data = await ctx.client.post("/api/process", {
+          input: readInputAsBase64(options.input),
+          operations: [{ op: "extract_frames", count: Number(options.count) }]
+        });
+        const localPath = saveProcessOutput(data, options.outputDir);
+        if (localPath) data.local_path = localPath;
+        printSuccess("process.extract_frames", data, ctx);
+      } catch (error2) {
+        printError("process.extract_frames", error2);
+      }
+    })
+  );
+  command.addCommand(
+    new Command5("remove-bg").description("Remove background from image(s) using rembg or ImageMagick fallback").requiredOption("--input <paths...>", "Input images (files or URLs)").option("--bg-color <color>", "Background color hint for fallback (e.g. white, black)").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const inputs = Array.isArray(options.input) ? options.input : [options.input];
+        const op = { op: "remove_bg" };
+        if (options.bgColor) op.bg_color = options.bgColor;
+        const data = await ctx.client.post("/api/process", {
+          inputs: inputs.map(readInputAsBase64),
+          operations: [op]
+        });
+        const localPath = saveProcessOutput(data, options.outputDir);
+        if (localPath) data.local_path = localPath;
+        printSuccess("process.remove_bg", data, ctx);
+      } catch (error2) {
+        printError("process.remove_bg", error2);
+      }
+    })
+  );
   return command;
 }
 
@@ -822,6 +1154,46 @@ async function saveProcess3dOutput(data, operation, outputDir, format) {
   writeFileSync4(filePath, buffer);
   return filePath;
 }
+function decodeBase64Payload(payload) {
+  const marker = ";base64,";
+  const index = payload.indexOf(marker);
+  const raw = index >= 0 ? payload.slice(index + marker.length) : payload;
+  return Buffer.from(raw, "base64");
+}
+async function saveRenderSpritesOutput(data, outputDir) {
+  const frames = Array.isArray(data.frames) ? data.frames : null;
+  if (!frames) {
+    return;
+  }
+  mkdirSync4(outputDir, { recursive: true });
+  const localPaths = [];
+  for (let index = 0; index < frames.length; index += 1) {
+    const frame = frames[index];
+    if (!isRecord2(frame)) {
+      continue;
+    }
+    const filename = typeof frame.filename === "string" ? frame.filename : `frame_${String(index).padStart(4, "0")}.png`;
+    if (typeof frame.image_base64 !== "string") {
+      continue;
+    }
+    const outputPath = join4(outputDir, filename);
+    writeFileSync4(outputPath, decodeBase64Payload(frame.image_base64));
+    delete frame.image_base64;
+    frame.local_path = outputPath;
+    localPaths.push(outputPath);
+  }
+  if (isRecord2(data.metadata)) {
+    const metadataPath = join4(outputDir, "metadata.json");
+    writeFileSync4(metadataPath, `${JSON.stringify(data.metadata, null, 2)}
+`);
+    data.local_metadata_path = metadataPath;
+  }
+  data.output_dir = outputDir;
+  data.local_paths = localPaths;
+}
+function isRecord2(value) {
+  return typeof value === "object" && value !== null && !Array.isArray(value);
+}
 function createProcess3dCommand() {
   const command = new Command6("process3d").description("3D model post-processing via Tripo pipeline");
   command.addCommand(
@@ -917,6 +1289,25 @@ function createProcess3dCommand() {
       }
     })
   );
+  command.addCommand(
+    new Command6("render-sprites").description("Render animated 3D model to 2D sprite frames via Blender").requiredOption("--task-id <id>", "Tripo task ID").option("--frame-count <n>", "Number of frames", "8").option("--resolution <n>", "Frame resolution in pixels", "64").option("--camera-angle <angle>", "Camera: front, side, iso, 3/4, top", "front").option("--directions <n>", "Rotation directions: 1, 4, 8", "1").option("--output-dir <dir>", "Output directory", ".").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const data = await ctx.client.post("/api/process3d", {
+          task_id: options.taskId,
+          operation: "render_sprites",
+          frame_count: Number(options.frameCount),
+          resolution: Number(options.resolution),
+          camera_angle: options.cameraAngle,
+          directions: Number(options.directions)
+        });
+        await saveRenderSpritesOutput(data, options.outputDir);
+        printSuccess("process3d.render_sprites", data, ctx);
+      } catch (error2) {
+        printError("process3d.render_sprites", error2);
+      }
+    })
+  );
   command.addCommand(
     new Command6("reduce").description("Reduce polygon count (high-poly to low-poly)").requiredOption("--task-id <id>", "Tripo task ID").option("--face-limit <n>", "Target face count").option("--quad", "Use quad topology").option("--output-dir <dir>", "Directory to save output", ".").action(async function(options) {
       try {
@@ -1103,8 +1494,105 @@ function createUploadCommand() {
   return command;
 }
 
+// src/commands/voice.ts
+import { existsSync as existsSync4, readFileSync as readFileSync3 } from "fs";
+import { extname } from "path";
+import { Command as Command9 } from "commander";
+function inferAudioMime(filePath) {
+  const extension = extname(filePath).toLowerCase();
+  const map = {
+    ".mp3": "audio/mpeg",
+    ".wav": "audio/wav",
+    ".pcm": "audio/pcm",
+    ".opus": "audio/opus",
+    ".ogg": "audio/ogg",
+    ".m4a": "audio/mp4",
+    ".aac": "audio/aac",
+    ".flac": "audio/flac"
+  };
+  return map[extension] ?? "application/octet-stream";
+}
+function readAudioAsBase64(filePath) {
+  if (!existsSync4(filePath)) {
+    throw configError(`Audio file not found: ${filePath}`);
+  }
+  const bytes = readFileSync3(filePath);
+  return {
+    audio_base64: bytes.toString("base64"),
+    audio_mime: inferAudioMime(filePath)
+  };
+}
+function withVoiceType(path, type) {
+  if (!type) {
+    return path;
+  }
+  const params = new URLSearchParams({ type });
+  return `${path}?${params.toString()}`;
+}
+function createVoiceCommand() {
+  const command = new Command9("voice").description("Manage Qwen3-TTS custom voices");
+  command.addCommand(
+    new Command9("clone").description("Clone a voice from an audio sample").requiredOption("--audio <path>", "Reference audio file path").requiredOption("--name <name>", "Name for the cloned voice").option("--target-model <model>", "Voice cloning model, e.g. qwen3-tts-vc-2026-01-22").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const audio = readAudioAsBase64(options.audio);
+        const body = {
+          ...audio,
+          name: options.name
+        };
+        if (options.targetModel) body.target_model = options.targetModel;
+        const data = await ctx.client.post("/api/voice/clone", body);
+        printSuccess("voice.clone", data, ctx);
+      } catch (error2) {
+        printError("voice.clone", error2);
+      }
+    })
+  );
+  command.addCommand(
+    new Command9("design").description("Create a synthetic voice from a text description").requiredOption("--prompt <text>", "Voice description prompt").requiredOption("--preview-text <text>", "Preview text for the generated sample").requiredOption("--name <name>", "Name for the designed voice").option("--target-model <model>", "Voice design model, e.g. qwen3-tts-vd-2026-01-26").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const body = {
+          voice_prompt: options.prompt,
+          preview_text: options.previewText,
+          name: options.name
+        };
+        if (options.targetModel) body.target_model = options.targetModel;
+        const data = await ctx.client.post("/api/voice/design", body);
+        printSuccess("voice.design", data, ctx);
+      } catch (error2) {
+        printError("voice.design", error2);
+      }
+    })
+  );
+  command.addCommand(
+    new Command9("list").description("List custom cloned or designed voices").option("--type <type>", "Voice type: vc or vd").action(async function(options) {
+      try {
+        const ctx = createContext(this);
+        const data = await ctx.client.get(withVoiceType("/api/voice/list", options.type));
+        printSuccess("voice.list", data, ctx);
+      } catch (error2) {
+        printError("voice.list", error2);
+      }
+    })
+  );
+  command.addCommand(
+    new Command9("delete").description("Delete a custom cloned or designed voice").argument("<voice-id>", "Voice ID to delete").option("--type <type>", "Voice type: vc or vd").action(async function(voiceId, options) {
+      try {
+        const ctx = createContext(this);
+        const path = withVoiceType(`/api/voice/${encodeURIComponent(voiceId)}`, options.type);
+        const data = await ctx.client.delete(path);
+        printSuccess("voice.delete", data, ctx);
+      } catch (error2) {
+        printError("voice.delete", error2);
+      }
+    })
+  );
+  return command;
+}
+
 // src/index.ts
-var program = new Command9().name("asset-gateway").description("Universal asset generation gateway CLI").version(CLI_VERSION).option(
+var program = new Command10().name("asset-gateway").description("Universal asset generation gateway CLI").version(CLI_VERSION).option(
   "--gateway-url <url>",
   `Gateway URL (default: $ASSET_GATEWAY_URL, auth config, or ${DEFAULT_GATEWAY_URL})`
 ).option("--token <token>", "API token for authentication").option("--human", "Human-readable output instead of JSON").option("--fields <fields>", "Comma-separated list of output fields");
@@ -1114,6 +1602,7 @@ program.addCommand(createProcessCommand());
 program.addCommand(createProcess3dCommand());
 program.addCommand(createProviderCommand());
 program.addCommand(createUploadCommand());
+program.addCommand(createVoiceCommand());
 program.addCommand(createJobCommand());
 program.addCommand(createDescribeCommand());
 await program.parseAsync(process.argv);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@doufunao123/asset-gateway",
-  "version": "0.11.0",
+  "version": "0.12.0",
   "description": "Universal asset generation gateway CLI",
   "type": "module",
   "bin": {
@@ -20,6 +20,7 @@
     "build": "tsup src/index.ts --format esm --dts --clean",
     "dev": "tsup src/index.ts --format esm --watch",
     "lint": "tsc --noEmit",
+    "test:process3d": "bash ../scripts/e2e-process3d-chain.sh",
     "prepublishOnly": "npm run build"
   },
   "engines": {