agentleFS
yyert123yyert123/miragea-skills/video-captions/SKILL.md

video-captions

Skill0 starsChanged 2 days agoMITView on GitHub

Caption any video or audio file in one command. Produces clean, readable SRT and WebVTT subtitles (whole words, 42-char lines, proper timing) and can burn them into an MP4, including big centred 1-3 word "pop" captions for TikTok, Reels and YouTube Shorts. Use when the user asks to add subtitles or captions to a video, make an SRT/VTT file, transcribe a talk or podcast into captions, or burn captions into a short-form video. English-first; speech recognition by Miragea.

  • Reads credentials
  • Installs packages

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Check MIRAGEA_API_KEY is set (never print it).

    line 25
  2. Distinctive:

    Run the script on the user's file:

    line 27
  3. Very specific:

    Read the JSON it prints (srt, vtt, video, cues) and give the user the paths.

    line 35
  4. Very specific:

    If the user wants wording fixed, edit the .srt text (keep timestamps) and re-run with --burn: the transcript is cached in …

    line 36
  5. Very specific:

    Only a 16 kHz mono audio track is uploaded (not the video), up to 90 minutes / 1 GB per file.

    line 52

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Video captions

One command turns a video (or audio) file into subtitles that are pleasant to read, and optionally into a captioned MP4. Raw speech-recognition segments often break mid-word ("cen" / "ts") and run too long; this skill rebuilds whole words and re-flows them into proper cues.

When to use

  • "Add captions / subtitles to this video", "make an SRT for my talk", "burn subtitles in".
  • "Make TikTok/Reels/Shorts style captions" -> use --style shorts --burn.
  • The user needs a .vtt for a website <track> or YouTube upload.

Steps

  1. Check MIRAGEA_API_KEY is set (never print it). If missing, tell the user to create a key at https://miragea.ai and export MIRAGEA_API_KEY=....
  2. Run the script on the user's file:

    python3 scripts/captions.py path/to/video.mp4                 # -> video.srt + video.vtt
    python3 scripts/captions.py path/to/video.mp4 --burn          # + video.captioned.mp4
    python3 scripts/captions.py path/to/short.mp4 --style shorts --burn   # pop captions, centred
  3. Read the JSON it prints (srt, vtt, video, cues) and give the user the paths.
  4. If the user wants wording fixed, edit the .srt text (keep timestamps) and re-run with --burn: the transcript is cached in <file>.transcript.json, so re-runs cost nothing.

Options

FlagDefaultWhat it does
--style standardyes2 lines x 42 chars, 1-7 s per cue (broadcast conventions)
--style shorts1-3 words per cue, UPPERCASE, centred when burned
--burnoffWrites <name>.captioned.mp4 (H.264, original audio copied)
--upperoffUppercase standard captions
--out-dir DIRnext to inputWhere outputs go
--forceoffIgnore cached transcript and re-transcribe

Cost and limits

Only a 16 kHz mono audio track is uploaded (not the video), up to 90 minutes / 1 GB per file. Miragea bills 3 points per second of audio; cached re-runs are free. Check live rates with GET https://miragea.ai/v1/capabilities.

Troubleshooting

  • ffmpeg not found -> install ffmpeg (brew install ffmpeg, sudo apt install ffmpeg, winget install ffmpeg).
  • 401 -> the key is wrong; 402 -> top up at miragea.ai; 429 -> wait and re-run (safe to retry).
  • Captions slightly early/late on music-heavy clips: re-run with --force after trimming intros.

See examples/ for sample output.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.