transcription-reader-marketplace

Transcription Reader

An AI agent skill for reading, parsing, and analyzing transcription and subtitle files. Strips timing metadata to produce compact, analysis-ready text — saving 40-90% of tokens depending on format.

Uses the open Agent Skills standard. Works with Claude, ChatGPT, Codex CLI, Cursor, Windsurf, Manus, and any other compatible tool.

Supported Formats

Extension	Format	Key Features
.stj, .stjson, .stj.json	STJ (Standard Transcription JSON)	Speaker names, language per segment, confidence scores, word-level timing
.vtt	WebVTT	Voice tags and inline speaker labels (Zoom compatible)
.srt	SubRip	Simple numbered blocks, optional inline speakers
.ass, .ssa	Advanced SubStation Alpha	Named dialogue events, style stripping

Installation

Claude Desktop

1. Click Customize in the sidebar
2. Click Browse Plugins
3. Go to the Personal tab and click +
4. Add: yaniv-golan/transcription-reader-skill

Claude Code (CLI)

/plugin marketplace add yaniv-golan/transcription-reader-skill

Cursor

1. Open Cursor Settings
2. Paste https://github.com/yaniv-golan/transcription-reader-skill into the Search or Paste Link box

Manus

1. Download transcription-reader.zip
2. Go to Settings → Skills
3. Click + Add → Upload
4. Upload the zip

ChatGPT

1. Download transcription-reader.zip
2. Go to Settings → Skills → New Skill → Upload from your Computer
3. Upload the zip

Codex CLI

Use the built-in skill installer:

$skill-installer https://github.com/yaniv-golan/transcription-reader-skill

Or install manually:

1. Download transcription-reader.zip
2. Extract the transcription-reader/ folder to ~/.codex/skills/

Other Tools (Windsurf, etc.)

Download transcription-reader.zip and extract the transcription-reader/ folder to:

• Project-level: .agents/skills/ in your project root
• User-level: ~/.agents/skills/

Usage

The skill auto-activates when you attach or reference a transcription file. You can also invoke it manually with /transcription-reader:

/transcription-reader summarize the first 20 minutes of the attached conversation

How It Works

The skill teaches your AI agent to:

1. Identify the format from the file extension
2. Decide whether to extract — small files can be read directly, large files benefit from the bundled extraction script (saves thousands of tokens)
3. Run the extraction script to produce clean, speaker-labeled text
4. Analyze based on your request — summarize, search, extract action items, identify speakers, etc.

Extraction Script

The bundled extract_transcript.py supports:

Usage: extract_transcript.py INPUT_FILE [options]

Options:
  --format FORMAT        Force format (auto-detected from extension if omitted)
  --output FILE          Write to file instead of stdout
  --output-format FMT    Output as 'text' (default) or 'jsonl' (one JSON object per line)
  --keep-timestamps      Include timestamps in output
  --merge-speakers       Merge consecutive segments from the same speaker
  --time-range RANGE     Extract only a time range (e.g., 10:00-20:00)
  --min-confidence N     Skip low-confidence segments (STJ only)
  --speakers-only NAME   Filter to a specific speaker
  --language LANG        Filter by language (STJ only)
  --list-speakers        List speakers found in the file
  --stats                Show transcript statistics