diff --git a/README.md b/README.md index e503701..397f588 100644 --- a/README.md +++ b/README.md @@ -6,7 +6,7 @@ An agent skill that helps plan video edits through iterative dialogue — transc ## Features -- **Audio transcription** — bundled [funasr-script](https://github.com/modelscope/FunASR) with Fun-ASR-Nano (high quality, Chinese-optimized) and SenseVoice (fast preview) models +- **Audio transcription** — bundled [funasr-script](https://github.com/modelscope/FunASR) with Fun-ASR-Nano (default, high quality, Chinese-optimized) and SenseVoice (fast preview) models - **On-demand frame extraction** — ffmpeg-based clip cutting (`-c copy`) + scene-change detection + uniform sampling, with hardware acceleration (CUDA/NVDEC preferred) - **Artifact indexing** — SQLite database tracks all transcriptions, clips, and frames to avoid duplicate processing across sessions - **Vision analysis guidance** — binary-search-style frame sampling strategy; works with any vision-capable model the agent's runtime provides @@ -31,13 +31,13 @@ This skill is compatible with the [open agent skills ecosystem](https://www.npmj ```bash # Install globally (available across all projects) -npx skills add git@git.nite07.com:nite/video-edit-planner-skill.git -g -y +npx skills add https://git.nite07.com/nite/video-edit-planner-skill.git -g -y # Install to specific agents -npx skills add git@git.nite07.com:nite/video-edit-planner-skill.git -g -a claude-code -y +npx skills add https://git.nite07.com/nite/video-edit-planner-skill.git -g -a claude-code -y # List available skills without installing -npx skills add git@git.nite07.com:nite/video-edit-planner-skill.git --list +npx skills add https://git.nite07.com/nite/video-edit-planner-skill.git --list ``` Supports 73+ agent frameworks including Claude Code, Codex, Cursor, OpenCode, and more. @@ -45,21 +45,11 @@ Supports 73+ agent frameworks including Claude Code, Codex, Cursor, OpenCode, an **Option 2: `git clone` (manual)** ```bash -git clone git@git.nite07.com:nite/video-edit-planner-skill.git ~/.hermes/skills/media/video-edit-planner +git clone https://git.nite07.com/nite/video-edit-planner-skill.git ~/.hermes/skills/media/video-edit-planner ``` For Hermes Agent specifically, clone into `~/.hermes/skills/media/`. For other agents, follow their respective skills directory conventions. -### First use - -The transcription scripts need a one-time dependency setup: - -```bash -cd ~/.hermes/skills/media/video-edit-planner/scripts/transcription && uv sync -``` - -This creates an isolated venv with `funasr`, `torch`, `torchaudio`. If you've used these packages via uv before, the cache is reused — first-time cost is only the link step. - ## Workflow ``` @@ -73,26 +63,6 @@ This creates an isolated venv with `funasr`, `torch`, `torchaudio`. If you've us 8. Iterate — user asks follow-ups, plan is refined ``` -## Stock Material Search (optional) - -When planning edits, users may need stock materials — stickers, GIFs, meme video clips, or B-roll. The skill provides guidance for finding these resources on demand. - -### Material sources - -| Source | URL | Content | Best for | -|---|---|---|---| -| 爱给网 | https://www.aigei.com/ | Chinese meme video clips (热梗, 42K+), free download | Chinese-language creators (B站/抖音) | -| B站素材酷 | https://cool.bilibili.com/ | Bilibili platform-native assets (video, audio, BGM, templates, stickers) | B站 creators; large Chinese meme collection | -| Tenor | https://tenor.com/ | GIFs + stickers, 30+ languages incl. Chinese, free API | Quick GIF/sticker search | -| GIPHY | https://giphy.com/ | Largest GIF library, 30+ languages, sticker channel | English/multilingual GIF search | -| Klipy | https://klipy.com/ | Localized GIF/sticker/clip API, supports Chinese | API-integrated search, localization | - -For Chinese-language users, 爱给网 and B站素材酷 are recommended first. - -### B站素材酷 search - -B站素材酷 has no built-in search. Use the [bcut-resource-search](https://git.nite07.com/nite/bcut-resource-search) Tampermonkey userscript to add search functionality, or let the agent query the API directly. - ## Project Structure ``` @@ -105,7 +75,7 @@ video-edit-planner/ │ │ ├── pyproject.toml │ │ ├── uv.lock │ │ ├── funasr_common.py # Shared: ffprobe, audio extraction, model runner -│ │ ├── funasr_nano.py # Fun-ASR-Nano entry point (high quality) +│ │ ├── funasr_nano.py # Fun-ASR-Nano entry point (default, high quality) │ │ ├── funasr_fast.py # SenseVoice entry point (fast preview) │ │ └── funasr_regular.py # Paraformer entry point (comparison) │ ├── frames/ diff --git a/README.zh.md b/README.zh.md index 40749b1..46e776d 100644 --- a/README.zh.md +++ b/README.zh.md @@ -6,7 +6,7 @@ ## 功能特性 -- **音频转录** — 内置 [funasr-script](https://github.com/modelscope/FunASR)，支持 Fun-ASR-Nano（高质量，中文优化）和 SenseVoice（快速预览）两种模型 +- **音频转录** — 内置 [funasr-script](https://github.com/modelscope/FunASR)，默认使用 Fun-ASR-Nano（高质量，中文优化），SenseVoice 可用于快速预览 - **按需抽帧** — 基于 ffmpeg 的片段剪辑（`-c copy`）+ 场景变化检测 + 均匀采样，优先使用硬件加速（CUDA/NVDEC） - **产物索引** — SQLite 数据库记录所有转录、剪辑片段和帧图片，避免跨会话重复处理 - **视觉分析指引** — 二分查找式帧采样策略；兼容 agent 运行时提供的任何视觉模型 @@ -31,13 +31,13 @@ ```bash # 全局安装（跨项目可用） -npx skills add git@git.nite07.com:nite/video-edit-planner-skill.git -g -y +npx skills add https://git.nite07.com/nite/video-edit-planner-skill.git -g -y # 安装到指定 agent -npx skills add git@git.nite07.com:nite/video-edit-planner-skill.git -g -a claude-code -y +npx skills add https://git.nite07.com/nite/video-edit-planner-skill.git -g -a claude-code -y # 列出可用 skill（不安装） -npx skills add git@git.nite07.com:nite/video-edit-planner-skill.git --list +npx skills add https://git.nite07.com/nite/video-edit-planner-skill.git --list ``` 支持 73+ agent 框架，包括 Claude Code、Codex、Cursor、OpenCode 等。 @@ -45,21 +45,11 @@ npx skills add git@git.nite07.com:nite/video-edit-planner-skill.git --list **方式二：`git clone`（手动）** ```bash -git clone git@git.nite07.com:nite/video-edit-planner-skill.git ~/.hermes/skills/media/video-edit-planner +git clone https://git.nite07.com/nite/video-edit-planner-skill.git ~/.hermes/skills/media/video-edit-planner ``` Hermes Agent 用户克隆到 `~/.hermes/skills/media/`。其他 agent 请遵循各自的 skills 目录约定。 -### 首次使用 - -转录脚本需要一次性依赖安装： - -```bash -cd ~/.hermes/skills/media/video-edit-planner/scripts/transcription && uv sync -``` - -这会创建一个隔离的 venv，包含 `funasr`、`torch`、`torchaudio`。如果你之前通过 uv 安装过这些包，缓存会被复用——首次成本仅为链接步骤。 - ## 工作流程 ``` @@ -73,26 +63,6 @@ cd ~/.hermes/skills/media/video-edit-planner/scripts/transcription && uv sync 8. 迭代——用户追问，规划逐步细化 ``` -## 素材搜索（可选） - -规划剪辑时，用户可能需要素材——贴纸、GIF、梗图视频片段、B-roll 等。skill 按需提供素材搜索指引。 - -### 素材来源 - -| 来源 | 网址 | 内容 | 适用场景 | -|---|---|---|---| -| 爱给网 | https://www.aigei.com/ | 中文梗图视频片段（热梗，4.2万+），免费下载 | 中文创作者（B站/抖音） | -| B站素材酷 | https://cool.bilibili.com/ | 哔哩哔哩平台素材（视频、音频、BGM、模板、贴纸） | B站创作者；大量中文梗图 | -| Tenor | https://tenor.com/ | GIF + 贴纸，支持 30+ 语言含中文，免费 API | 快速 GIF/贴纸搜索 | -| GIPHY | https://giphy.com/ | 最大 GIF 库，30+ 语言，贴纸频道 | 英文/多语言 GIF 搜索 | -| Klipy | https://klipy.com/ | 本地化 GIF/贴纸/片段 API，支持中文 | API 集成搜索，本地化 | - -中文用户优先推荐爱给网和B站素材酷。 - -### B站素材酷搜索 - -B站素材酷没有内置搜索功能。可以使用 [bcut-resource-search](https://git.nite07.com/nite/bcut-resource-search) 油猴脚本添加搜索功能，或让 agent 直接查询 API。 - ## 目录结构 ``` @@ -105,7 +75,7 @@ video-edit-planner/ │ │ ├── pyproject.toml │ │ ├── uv.lock │ │ ├── funasr_common.py # 共享：ffprobe、音频提取、模型运行 -│ │ ├── funasr_nano.py # Fun-ASR-Nano 入口（高质量） +│ │ ├── funasr_nano.py # Fun-ASR-Nano 入口（默认，高质量） │ │ ├── funasr_fast.py # SenseVoice 入口（快速预览） │ │ └── funasr_regular.py # Paraformer 入口（对比用） │ ├── frames/ diff --git a/SKILL.md b/SKILL.md index 9bb140c..42a0ee5 100644 --- a/SKILL.md +++ b/SKILL.md @@ -72,15 +72,15 @@ User requirements are often vague in the first pass — that's expected. The pla ### Step 4 — Transcribe audio -Use the bundled `funasr-script` in `scripts/transcription/`. +Use the bundled `funasr-script` in `scripts/transcription/`. **Prefer Fun-ASR-Nano** as the default transcription model — it has higher accuracy and sensitivity on Chinese audio. Use SenseVoice only for fast previews. **First run** requires `uv sync` (installs funasr + torch, may take several minutes). Subsequent runs reuse the cached venv. Notify the user before first-time setup. ```bash -# Default: Fun-ASR-Nano (high quality, Chinese-optimized) +# Recommended default: Fun-ASR-Nano (high quality, Chinese-optimized) uv run --directory SKILL_DIR/scripts/transcription funasr-nano VIDEO_PATH --track TRACK_INDEX --output-dir VIDEO_DIR -# Fast preview: SenseVoice +# Fast preview only: SenseVoice uv run --directory SKILL_DIR/scripts/transcription funasr-fast VIDEO_PATH --track TRACK_INDEX --output-dir VIDEO_DIR # List tracks