AgenticIntros

Branded, AI-generated video intros for Profullstack and related projects, plus their transcripts.

The <video> player renders on GitHub only when intro.mp4 is tracked in the repo. It's currently gitignored (see Note on media files); the clip is ~3 MB, so it can be force-added if you want it to play inline.

Contents

• intro.* — the rendered Profullstack intro in multiple formats (mp3, mp4, wav, webm).
• Profullstack_Logo_Intro_7s_blank_fadeblack.mp4 — 7s logo intro with a black fade-out.
• input.* — source recording in multiple formats.
• transcript/ — Whisper-style transcripts of input in several formats:
  • input.json — segments with timestamps and metadata
  • input.srt / input.vtt — subtitle formats
  • input.tsv — tab-separated start/end/text
  • input.txt — plain text transcript

Build

build.sh regenerates the media artifacts from their sources (needs ffmpeg, ffprobe, and whisper):

./build.sh intro       # render intro.{mp4,webm,wav,mp3} from the logo clip
./build.sh transcribe  # extract input.wav + write transcript/{json,srt,tsv,txt,vtt}
./build.sh all         # both (default)

Override defaults via env vars, e.g. INPUT_SRC=talk.webm WHISPER_MODEL=medium ./build.sh transcribe.

Note on media files

The large media files (*.mp4, *.webm, *.wav, *.mp3, …) are gitignored because they exceed GitHub's 100MB per-file limit. Only the text transcripts are tracked in this repo. Store the binaries in object storage or Git LFS if they need to be versioned.

License

MIT © 2026 Profullstack, Inc.