agentsclimarketplace

Caption conversation

Skill sumitaich1998/jarvisvr/skills/communication/caption-conversation

An AI agentic operating system for mixed reality on the Meta Quest 3 — your own J.A.R.V.I.S. Multi-agent orchestration, multimodal perception (sight/hearing/gaze), 42 holographic widgets, 20 LLM providers.

Install
npx -y skills add sumitaich1998/jarvisvr --skill caption-conversation

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Show live, rolling captions of speech Jarvis hears — for accessibility, noisy rooms, or meetings — with optional translation. Use for "caption this", "turn on subtitles", "caption what they're saying", or "subtitle this meeting". Triggers: caption, captions, subtitles, transcribe live, what are they saying, accessibility, hard of hearing, meeting transcript.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

2.7 KB, as published. Nobody here has run it

Caption Conversation

Render a live_caption panel that scrolls speech Jarvis hears in real time, tagged by speaker, optionally translated.

Steps

  1. Ensure ambient audio is on: perception.request{stream:"ambient_audio", action:"start", reason:"user asked for live captions"}.
  2. Open show_live_caption with an empty/seed lines array and the right speaker and max_lines.
  3. Per audio window: append perception.audio_scene.ambient_transcript to lines (newest last), set speaker (user|other|jarvis|unknown), and holo.update. Trim to max_lines.
  4. Translate (optional): if a target language is requested, translate each line, set translated:true and language (pair with live-translate).
  5. Stop captions and the audio stream when the user is done.

Output

live_caption (show_live_caption, props per registry.json):

{ "widget_type": "live_caption",
  "transform": { "anchor": "head", "position": [0.0,-0.35,1.0], "billboard": true },
  "props": { "lines": [ "Hello there.", "How can I help you today?" ],
             "speaker": "other", "max_lines": 3, "language": "en", "translated": false },
  "interactions": ["grab","tap","resize"] }

Update as speech arrives (holo.update):

{ "object_id": "O_caption",
  "props": { "lines": [ "How can I help you today?", "I'd like the window seat." ], "speaker": "other" } }

Edge cases

  • Multiple speakers → tag turns by speaker; "unknown" when diarization is uncertain.
  • Background noise / no speech → keep the panel idle; don't caption music or ambient sounds (that's perception identify_sound).
  • Long sessions → cap visible lines (max_lines ≤ 10); the full transcript can be saved as a note.
  • Translated captions → set translated:true; consider a two-line layout (original + translation) via live-translate.
  • Privacy → captioning others' speech is sensitive; it's user-initiated and perception.state reflects active capture. Stop promptly when asked.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.