AI Skill Hub 推荐使用:pavo-engine-py 是一款优质的Agent工作流。AI 综合评分 6.0 分,在同类工具中表现稳健。如果你正在寻找可靠的Agent工作流解决方案,这是一个值得深入了解的选择。
pavo-engine-py 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
pavo-engine-py 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:pip 安装(推荐)
pip install pavo-engine-py
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install pavo-engine-py
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/sonnhfit/pavo-engine-py
cd pavo-engine-py
pip install -e .
# 验证安装
python -c "import pavo_engine_py; print('安装成功')"
# 命令行使用
pavo-engine-py --help
# 基本用法
pavo-engine-py input_file -o output_file
# Python 代码中调用
import pavo_engine_py
# 示例
result = pavo_engine_py.process("input")
print(result)
# pavo-engine-py 配置文件示例(config.yml) app: name: "pavo-engine-py" debug: false log_level: "INFO" # 运行时指定配置文件 pavo-engine-py --config config.yml # 或通过环境变量配置 export PAVO_ENGINE_PY_API_KEY="your-key" export PAVO_ENGINE_PY_OUTPUT_DIR="./output"
Pavo Engine is an intelligent AI agent for automated video editing that transforms natural language prompts into professionally edited videos. Built with a modular 4-layer architecture, it combines computer vision, speech processing, and timeline automation to create dynamic video content.
openai-whisper>=20230314 # Speech recognition ultralytics # Object detection scenedetect[opencv]>=0.6.2 # Scene detection ```
pip install -r requirements.txt
```bash
ffmpeg-python==0.2.0 # FFmpeg integration requests==2.32.3 # HTTP requests retrying==1.3.4 # Retry logic tqdm==4.66.4 # Progress bars pillow==10.4.0 # Image processing
```bash
```bash
pip install -e .
python setup.py sdist bdist_wheel ```
```python from pavo import render_video
Check the docs/example/ directory for comprehensive examples:
VIDEO_CONVERTER_EXAMPLE.py - Video format conversionAUDIO_CONVERTER_EXAMPLE.py - Audio processingSPEECH_TRANSCRIBER_EXAMPLE.py - Speech-to-textSPEAKER_DIARIZATION_EXAMPLE.py - Speaker identificationOBJECT_DETECTION_EXAMPLE.py - Object recognitionSCENE_DETECTION_EXAMPLE.py - Scene change detectionSave the following as timeline.json and pass its path to render_video:
{
"timeline": {
"n_frames": 75,
"background": "#1a1a2e",
"soundtrack": {
"src": "path/to/background_music.mp3",
"effect": "fadeOut"
},
"tracks": [
{
"track_id": 0,
"strips": [
{
"asset": {
"type": "image",
"src": "path/to/intro.jpg"
},
"start": 0,
"video_start_frame": 0,
"length": 25,
"effect": "zoomIn",
"transition": {"in": "fade", "out": "fade"}
},
{
"asset": {
"type": "video",
"src": "path/to/clip.mp4"
},
"start": 25,
"video_start_frame": 0,
"length": 50,
"effect": null,
"transition": {"in": "fade", "out": "fade"}
}
]
}
]
},
"output": {
"format": "mp4",
"fps": 25,
"width": 1280,
"height": 720
}
}
Enable the audio_ducking flag in the output section to automatically lower the soundtrack volume whenever speech is detected in video clips:
{
"timeline": {
"n_frames": 75,
"background": "#1a1a2e",
"soundtrack": {
"src": "path/to/background_music.mp3"
},
"tracks": [
{
"track_id": 0,
"strips": [
{
"asset": {
"type": "video",
"src": "path/to/narration.mp4"
},
"start": 0,
"video_start_frame": 0,
"length": 75,
"effect": null,
"transition": {}
}
]
}
]
},
"output": {
"format": "mp4",
"fps": 25,
"width": 1280,
"height": 720,
"audio_ducking": true,
"ducking_reduction_db": 10
}
}
from pavo import render_video
render_video('timeline.json', 'output/video.mp4')
output field | Type | Default | Description |
|---|---|---|---|
audio_ducking | bool | false | Enable automatic volume ducking of soundtrack during detected speech segments. |
ducking_reduction_db | float | 10.0 | Volume reduction (in dB) applied to the soundtrack during speech. Requires openai-whisper. |
Specify optional trim_start / trim_end (seconds) or trim_start_frame / trim_end_frame (integer frame numbers) inside a video asset to use only a sub-segment of the source file. If only one boundary is given the other defaults to the natural start/end of the source.
{
"timeline": {
"n_frames": 50,
"background": "#000000",
"tracks": [
{
"track_id": 0,
"strips": [
{
"asset": {
"type": "video",
"src": "path/to/long_clip.mp4",
"trim_start": 5.0,
"trim_end": 15.0
},
"start": 0,
"length": 50
}
]
}
]
},
"output": {
"format": "mp4",
"fps": 25,
"width": 1280,
"height": 720
}
}
Frame-based trimming example (equivalent to the above at 25 fps):
{
"asset": {
"type": "video",
"src": "path/to/long_clip.mp4",
"trim_start_frame": 125,
"trim_end_frame": 375
}
}
asset field | Type | Default | Description |
|---|---|---|---|
trim_start | float | null | Start of the used segment in seconds (≥ 0). Video only. |
trim_end | float | null | End of the used segment in seconds (> 0). Video only. |
trim_start_frame | int | null | Start of the used segment in frames (≥ 0). Video only. Cannot be combined with trim_start. |
trim_end_frame | int | null | End of the used segment in frames (≥ 1). Video only. Cannot be combined with trim_end. |
render_video raises descriptive exceptions for common mistakes:
from pavo import render_video
try:
render_video('timeline.json', 'output/video.mp4')
except FileNotFoundError as exc:
print(f"JSON file missing: {exc}")
except ValueError as exc:
print(f"Invalid timeline JSON: {exc}")
converter.convert( 'input.mov', 'output.mp4', scale='1920:1080', fps=30, bitrate='5000k' ) ```
boto3==1.26.109 # AWS S3 botocore==1.29.109 # AWS core
python -m pytest tests/test_render.py
该工具未明确声明开源协议,商业使用前请联系原作者确认授权范围,避免侵权风险。
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
总体来看,pavo-engine-py 是一款质量良好的Agent工作流,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | pavo-engine-py |
| 原始描述 | 开源AI工作流:Pavo Engine - AI-Powered Video Editing Agent。⭐6 · Python |
| Topics | workflowaiai-agent-video-editorai-video-editorjson-to-videollm-video-editorpython |
| GitHub | https://github.com/sonnhfit/pavo-engine-py |
| 语言 | Python |
收录时间:2026-05-22 · 更新时间:2026-05-22 · License:未公布 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端