Turn your own video and matching SRT into source-linked vertical clips: MP4s, editable subtitles, a review page and original time ranges. Render locally with Python and FFmpeg, without a paid video API or per-render service fee. No transcription, URL download or publishing. Review context, caption alignment and audio before posting.
VideoDB from affaan-m’s ECC project. Use it to ingest, understand, search, edit, and monitor video or audio from files, URLs, live streams, or desktop sessions. Build visual and speech indexes, find moments by timestamp, generate clips and subtitles, transcode media, add overlays or audio, and create alerts for live events. Powered by the VideoDB SDK and API.
Programmatic access to Google NotebookLM via the nlm CLI — manages notebooks, sources, queries, and artifacts with citation-quality source titles.
Structure project docs as modular memory + reasoning files so AI agents load only what they need — cutting token usage by 70–96% per query. Grounded in Anthropic, LangChain, and Crawl4AI's production llm.txt research. Implements write/select/compress/isolate context engineering strategies.
LLMエージェントがセッションをまたいで記憶を保持するための永続メモリ層。抽象基底クラス MemoryStore ＋ LocalJSONStore 実装で、保存・取得・検索を再利用可能に。Karpathy の「LLM OS（記憶＝ディスク層）」着想。バックエンド差し替え可・Python標準ライブラリのみ・LLM呼び出しゼロ・ユニットテスト同梱。
ブランド調査、ライブ配信のアーカイブを検証するジャーナリスト、長い動画を最後まで見ずに要点を抽出したい人向けのFacebook動画文字起こしツールです。Watch動画、ライブ配信のアーカイブ、Reels、ページ投稿などの公開リンクを貼るだけで、逐語文字起こしとAI要約を生成します。プレーンテキスト、タイムスタンプ付きMarkdown、SRT字幕に対応。文字起こしは元の言語を保持し、要約はチャット言語に自動翻訳されます。
My 82M+ view YouTube video did not work because every frame was beautiful; it worked because people kept watching. After growing to nearly 800K YouTube subscribers and over 1.19M followers across platforms, with dozens of 10M+ and 1M+ videos, I learned to judge clips by retention, not just style. This AI Skill applies that judgment to short AI videos, shaping the first frame, pacing, motion, visual rhythm, and ending so viewers have a reason to stay to the last frame.