@plusonelabs/cue 0.0.94 → 0.0.95

package/dist/skills/seeddance/SKILL.md

@@ -0,0 +1,81 @@
+---
+name: seeddance
+description: Generate videos using ByteDance Doubao SeedAnce API. Use when user asks to generate videos with SeedAnce, Doubao, or ByteDance video models. Supports text-to-video, image-to-video, first/last frame interpolation, and audio generation.
+category: media
+type: hybrid
+env:
+  - ARK_API_KEY
+metadata:
+  short-description: Generate videos with SeedAnce
+  scope: first-party
+---
+
+Generate videos using ByteDance's Doubao SeedAnce API (Volcengine Ark).
+
+## Requirements
+
+- Python 3.10+, `httpx` package
+- `ARK_API_KEY` environment variable (or configured via `cue skill env set`)
+
+## Usage
+
+```bash
+# Text-to-video
+python3 <skill-dir>/generate.py "A field of white daisies under blue sky, camera slowly zooms in"
+
+# Image-to-video (animate a reference image)
+python3 <skill-dir>/generate.py "Girl opens her eyes and looks at camera" --image photo.jpg
+
+# First + last frame interpolation
+python3 <skill-dir>/generate.py "Girl says cheese, 360 orbit camera" --first-frame start.jpg --last-frame end.jpg
+
+# With audio generation
+python3 <skill-dir>/generate.py "Wind blowing through hair" --image photo.jpg --audio
+
+# Custom output path
+python3 <skill-dir>/generate.py "A cat walking" --output cat.mp4
+
+# Aspect ratios
+python3 <skill-dir>/generate.py "prompt" --ratio 16:9      # Landscape (default)
+python3 <skill-dir>/generate.py "prompt" --ratio 9:16      # Vertical
+python3 <skill-dir>/generate.py "prompt" --ratio 1:1       # Square
+python3 <skill-dir>/generate.py "prompt" --ratio adaptive   # Match input image
+
+# Duration (seconds)
+python3 <skill-dir>/generate.py "prompt" --duration 5      # 5 seconds (default)
+python3 <skill-dir>/generate.py "prompt" --duration 10     # 10 seconds
+
+# Use 1.0 model (1080p text-to-video, no image input)
+python3 <skill-dir>/generate.py "prompt" --v1
+```
+
+**Generation Time:** Typically 30-90 seconds. The script polls automatically and downloads when complete.
+
+## Models
+
+| Model            | ID                               | Resolution | Features                                                         |
+| ---------------- | -------------------------------- | ---------- | ---------------------------------------------------------------- |
+| SeedAnce 1.5 Pro | `doubao-seedance-1-5-pro-251215` | 720p       | Text-to-video, image-to-video, first/last frame, audio (default) |
+| SeedAnce 1.0 Pro | `doubao-seedance-1-0-pro-250528` | 1080p      | Text-to-video only                                               |
+
+## Combining with Image Generation
+
+Generate first frame with nano-banana, then animate with seeddance:
+
+```bash
+# 1. Generate image
+python3 <skills-dir>/nano-banana/generate.py "A girl holding a fox in a forest" --output fox_girl.png
+
+# 2. Animate the image
+python3 <skills-dir>/seeddance/generate.py "Girl opens eyes, looks at camera, wind blows hair" --image fox_girl.png --audio
+```
+
+## Troubleshooting
+
+| Error                  | Solution                                      |
+| ---------------------- | --------------------------------------------- |
+| "ARK_API_KEY not set"  | Export env var: `export ARK_API_KEY=your-key` |
+| "API error: 429"       | Rate limited - wait and retry                 |
+| "API error: 400"       | Check prompt or image URL format              |
+| "Generation timed out" | Try shorter duration or simpler prompt        |
+| "Generation failed"    | Check API status or try a different prompt    |

package/dist/skills/seeddance/generate.py

@@ -0,0 +1,303 @@
+#!/usr/bin/env python3
+"""
+SeedAnce - Video generation using ByteDance Doubao SeedAnce API.
+
+Usage:
+    python generate.py "A sunset over mountains"
+    python generate.py "prompt" --output video.mp4
+    python generate.py "prompt" --image first_frame.jpg
+    python generate.py "prompt" --first-frame start.jpg --last-frame end.jpg
+    python generate.py "prompt" --image photo.jpg --audio
+    python generate.py "prompt" --v1  # Use 1.0 model (1080p, text-only)
+"""
+
+import argparse
+import os
+import sys
+import time
+from datetime import datetime
+from pathlib import Path
+
+sys.path.insert(0, os.path.join(os.path.dirname(os.path.abspath(__file__)), '..', '_lib'))
+from credentials import get_credential
+
+try:
+    import httpx
+except ImportError:
+    print("Error: httpx not installed. Run: pip install httpx")
+    sys.exit(1)
+
+
+API_URL = "https://ark.cn-beijing.volces.com/api/v3/contents/generations/tasks"
+MODEL_V15 = "doubao-seedance-1-5-pro-251215"
+MODEL_V10 = "doubao-seedance-1-0-pro-250528"
+TIMEOUT = 30.0
+POLL_INTERVAL = 5
+MAX_POLL_TIME = 300
+SKILL_DIR = Path(__file__).parent
+ARCHIVE_DIR = SKILL_DIR / "_archive"
+
+
+def create_task(
+    prompt: str,
+    model: str = MODEL_V15,
+    ratio: str = "16:9",
+    duration: int = 5,
+    image_url: str | None = None,
+    first_frame_url: str | None = None,
+    last_frame_url: str | None = None,
+    generate_audio: bool = False,
+) -> str:
+    api_key = get_credential("ARK_API_KEY", "seeddance")
+
+    content: list[dict] = [{"type": "text", "text": prompt}]
+
+    if image_url:
+        content.append({
+            "type": "image_url",
+            "image_url": {"url": image_url},
+        })
+    if first_frame_url:
+        content.append({
+            "type": "image_url",
+            "image_url": {"url": first_frame_url},
+            "role": "first_frame",
+        })
+    if last_frame_url:
+        content.append({
+            "type": "image_url",
+            "image_url": {"url": last_frame_url},
+            "role": "last_frame",
+        })
+
+    body: dict = {
+        "model": model,
+        "content": content,
+        "ratio": ratio,
+        "duration": duration,
+        "watermark": False,
+    }
+    if generate_audio:
+        body["generate_audio"] = True
+
+    print(f"Starting video generation with {model}...")
+    print(f"Prompt: {prompt[:100]}{'...' if len(prompt) > 100 else ''}")
+
+    with httpx.Client(timeout=TIMEOUT) as client:
+        response = client.post(
+            API_URL,
+            json=body,
+            headers={
+                "Content-Type": "application/json",
+                "Authorization": f"Bearer {api_key}",
+            },
+        )
+
+        if response.status_code != 200:
+            print(f"Error: API returned {response.status_code}")
+            print(response.text[:500])
+            sys.exit(1)
+
+        data = response.json()
+
+    if "error" in data:
+        print(f"Error: {data['error'].get('message', data['error'])}")
+        sys.exit(1)
+
+    task_id = data.get("id")
+    if not task_id:
+        print("Error: No task ID in response")
+        print(data)
+        sys.exit(1)
+
+    return task_id
+
+
+def poll_task(task_id: str) -> tuple[bool, str | None, str | None]:
+    api_key = get_credential("ARK_API_KEY", "seeddance")
+    url = f"{API_URL}/{task_id}"
+
+    with httpx.Client(timeout=TIMEOUT) as client:
+        response = client.get(
+            url,
+            headers={"Authorization": f"Bearer {api_key}"},
+        )
+
+        if response.status_code != 200:
+            return False, None, f"Poll error: {response.status_code}"
+
+        data = response.json()
+
+    status = data.get("status")
+
+    if status == "succeeded":
+        video_url = data.get("content", {}).get("video_url")
+        if video_url:
+            resolution = data.get("resolution", "unknown")
+            ratio = data.get("ratio", "unknown")
+            duration = data.get("duration", "unknown")
+            print(f"  Completed: {resolution}, {ratio}, {duration}s")
+            return True, video_url, None
+        return True, None, "No video URL in response"
+
+    if status == "failed":
+        error = data.get("error", {})
+        msg = error.get("message", "Unknown error") if isinstance(error, dict) else str(error)
+        return True, None, msg
+
+    return False, None, None
+
+
+def download_video(video_url: str) -> bytes:
+    print("Downloading video...")
+
+    with httpx.Client(timeout=120.0, follow_redirects=True) as client:
+        response = client.get(video_url)
+
+        if response.status_code != 200:
+            print(f"Error: Download failed with {response.status_code}")
+            sys.exit(1)
+
+        return response.content
+
+
+def generate_video(
+    prompt: str,
+    model: str = MODEL_V15,
+    ratio: str = "16:9",
+    duration: int = 5,
+    image_url: str | None = None,
+    first_frame_url: str | None = None,
+    last_frame_url: str | None = None,
+    generate_audio: bool = False,
+) -> bytes:
+    task_id = create_task(
+        prompt=prompt,
+        model=model,
+        ratio=ratio,
+        duration=duration,
+        image_url=image_url,
+        first_frame_url=first_frame_url,
+        last_frame_url=last_frame_url,
+        generate_audio=generate_audio,
+    )
+
+    print(f"Task: {task_id}")
+    print("Polling for completion...")
+
+    start_time = time.time()
+    while time.time() - start_time < MAX_POLL_TIME:
+        done, video_url, error = poll_task(task_id)
+
+        if done:
+            if error:
+                print(f"Error: {error}")
+                sys.exit(1)
+            if video_url:
+                return download_video(video_url)
+            print("Error: No video URL in completed task")
+            sys.exit(1)
+
+        elapsed = int(time.time() - start_time)
+        print(f"  Still processing... ({elapsed}s elapsed)")
+        time.sleep(POLL_INTERVAL)
+
+    print(f"Error: Generation timed out after {MAX_POLL_TIME}s")
+    sys.exit(1)
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        description="Generate videos using ByteDance Doubao SeedAnce API",
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog="""
+Examples:
+  %(prog)s "A field of daisies under blue sky"
+  %(prog)s "prompt" --output video.mp4
+  %(prog)s "Animate this scene" --image https://example.com/photo.jpg
+  %(prog)s "Smooth transition" --first-frame https://example.com/start.jpg --last-frame https://example.com/end.jpg
+  %(prog)s "Wind blowing" --image https://example.com/photo.jpg --audio
+  %(prog)s "prompt" --v1  # 1080p text-to-video
+        """,
+    )
+    parser.add_argument("prompt", help="Video generation prompt")
+    parser.add_argument(
+        "--output", "-o",
+        help="Output file path (default: generated_<timestamp>.mp4)",
+    )
+    parser.add_argument(
+        "--image", "-i",
+        help="Input image URL for image-to-video (v1.5 only)",
+    )
+    parser.add_argument(
+        "--first-frame",
+        help="First frame image URL for interpolation (v1.5 only)",
+    )
+    parser.add_argument(
+        "--last-frame",
+        help="Last frame image URL for interpolation (v1.5 only, requires --first-frame)",
+    )
+    parser.add_argument(
+        "--audio", "-a",
+        action="store_true",
+        help="Generate audio along with video (v1.5 only)",
+    )
+    parser.add_argument(
+        "--ratio", "-r",
+        default="16:9",
+        choices=["16:9", "9:16", "1:1", "adaptive"],
+        help="Aspect ratio (default: 16:9, use 'adaptive' with image input)",
+    )
+    parser.add_argument(
+        "--duration", "-d",
+        type=int,
+        default=5,
+        help="Video duration in seconds (default: 5)",
+    )
+    parser.add_argument(
+        "--v1",
+        action="store_true",
+        help=f"Use v1.0 model ({MODEL_V10}) — 1080p text-to-video only",
+    )
+
+    args = parser.parse_args()
+
+    model = MODEL_V10 if args.v1 else MODEL_V15
+
+    if args.v1 and (args.image or args.first_frame or args.last_frame or args.audio):
+        print("Error: --v1 model only supports text-to-video (no --image, --first-frame, --last-frame, --audio)")
+        sys.exit(1)
+
+    if args.last_frame and not args.first_frame:
+        print("Error: --last-frame requires --first-frame")
+        sys.exit(1)
+
+    if args.image and args.first_frame:
+        print("Error: use --image for image-to-video OR --first-frame/--last-frame for interpolation, not both")
+        sys.exit(1)
+
+    video_bytes = generate_video(
+        prompt=args.prompt,
+        model=model,
+        ratio=args.ratio,
+        duration=args.duration,
+        image_url=args.image,
+        first_frame_url=args.first_frame,
+        last_frame_url=args.last_frame,
+        generate_audio=args.audio,
+    )
+
+    if args.output:
+        output_path = Path(args.output)
+    else:
+        ARCHIVE_DIR.mkdir(exist_ok=True)
+        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+        output_path = ARCHIVE_DIR / f"generated_{timestamp}.mp4"
+
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+    output_path.write_bytes(video_bytes)
+    print(f"Saved: {output_path} ({len(video_bytes):,} bytes)")
+
+
+if __name__ == "__main__":
+    main()

package/dist/skills/seedream/SKILL.md

@@ -0,0 +1,86 @@
+---
+name: seedream
+description: Generate images using ByteDance Doubao Seedream API. Use when user asks to generate images with Seedream, Doubao, or ByteDance image models. Supports text-to-image, image-to-image, multi-image input, and sequential image generation.
+category: media
+type: hybrid
+env:
+  - ARK_API_KEY
+metadata:
+  short-description: Generate images with Seedream
+  scope: first-party
+---
+
+Generate images using ByteDance's Doubao Seedream API (Volcengine Ark).
+
+## Requirements
+
+- Python 3.10+, `httpx` package
+- `ARK_API_KEY` environment variable (or configured via `cue skill env set`)
+
+## Usage
+
+```bash
+# Text-to-image
+python3 <skill-dir>/generate.py "A cute cat sitting on a windowsill, warm sunlight"
+
+# Custom output path
+python3 <skill-dir>/generate.py "A sunset over mountains" --output sunset.png
+
+# Image-to-image (reference image + prompt)
+python3 <skill-dir>/generate.py "Make it look like a watercolor painting" --image photo.jpg
+
+# Image editing (seededit model)
+python3 <skill-dir>/generate.py "Remove the background" --image photo.jpg --edit
+
+# Multiple reference images (up to 14)
+python3 <skill-dir>/generate.py "Combine these styles" --image a.jpg --image b.jpg --image c.jpg
+
+# Size options
+python3 <skill-dir>/generate.py "prompt" --size 2K          # 2K resolution (default)
+python3 <skill-dir>/generate.py "prompt" --size 4K          # 4K resolution
+python3 <skill-dir>/generate.py "prompt" --size 2560x1440   # Custom WxH
+
+# Sequential image generation (multiple related images)
+python3 <skill-dir>/generate.py "A four-panel comic about a cat" --sequential
+python3 <skill-dir>/generate.py "prompt" --sequential --max-images 4
+
+# Model selection
+python3 <skill-dir>/generate.py "prompt"                     # Seedream 5.0 (default)
+python3 <skill-dir>/generate.py "prompt" --v45               # Seedream 4.5
+python3 <skill-dir>/generate.py "prompt" --v4                # Seedream 4.0
+python3 <skill-dir>/generate.py "prompt" --v3                # Seedream 3.0 (text-only)
+python3 <skill-dir>/generate.py "prompt" --image x.jpg --edit  # SeedEdit 3.0
+```
+
+**Generation Time:** Typically 5-15 seconds.
+
+## Models
+
+| Model        | ID                               | Features                                                      |
+| ------------ | -------------------------------- | ------------------------------------------------------------- |
+| Seedream 5.0 | `doubao-seedream-5-0-260128`     | Text/image-to-image, multi-image, sequential, 2K/4K (default) |
+| Seedream 4.5 | `doubao-seedream-4-5-251128`     | Text/image-to-image, multi-image, sequential, 2K/4K           |
+| Seedream 4.0 | `doubao-seedream-4-0-250828`     | Text/image-to-image, multi-image, sequential, 1K/2K/4K        |
+| Seedream 3.0 | `doubao-seedream-3-0-t2i-250415` | Text-to-image only                                            |
+| SeedEdit 3.0 | `doubao-seededit-3-0-i2i-250628` | Image editing (single image + prompt)                         |
+
+## Combining with SeedAnce
+
+Generate an image with seedream, then animate it with seeddance:
+
+```bash
+# 1. Generate image
+python3 <skills-dir>/seedream/generate.py "A girl holding a fox in a snowy forest" --output fox_girl.png
+
+# 2. Animate with seeddance (pass URL or upload first)
+python3 <skills-dir>/seeddance/generate.py "Girl opens eyes, wind blows hair" --image fox_girl.png --audio
+```
+
+## Troubleshooting
+
+| Error                      | Solution                                      |
+| -------------------------- | --------------------------------------------- |
+| "ARK_API_KEY not set"      | Export env var: `export ARK_API_KEY=your-key` |
+| "API error: 429"           | Rate limited - wait and retry                 |
+| "API error: 400"           | Check prompt or image format                  |
+| "Content policy violation" | Modify prompt to avoid restricted content     |