AWS Caption Editor (Post-Processing Tool)

I built a Vue.js tool to post-process AWS Transcribe JSON output into editable captions. It displays transcripts with confidence-based highlighting, allows word edits, and exports VTT/SRT files. Useful for refining AI-generated captions for accuracy and accessibility in videos.
Problem / Context
AWS Transcribe provides JSON with timings and confidence scores, but editing requires manual processing. Needed a tool to visualize confidence, edit low-confidence words, and generate standard caption formats.
Constraints / Goals
- Load AWS JSON: parse items, alternatives, confidence.
- Visualize: highlight low-confidence words.
- Edit: click to modify words.
- Export: VTT/SRT with custom line lengths.
- Stats: show confidence metrics.
Core Approach
Vue app loads JSON, computes transcript with highlighting, generates VTT from timings.
Loading and Parsing
Assumes awsjson from file input.
Transcript Display
<template v-for="(item, index) in awsjson.results.items">
<template v-if="item.alternatives[0].content != '' && Number(item.alternatives[0].confidence) >= Number(slidercurrent)">
{{item.alternatives[0].content}}
</template>
<span v-on:click="editword(index)" v-if="Number(item.alternatives[0].confidence) < Number(slidercurrent)" class="bg-warning">
{{item.alternatives[0].content}}
</span>
</template>
Confidence Filtering
Slider to set threshold; highlights below it.
Editing Words
Click low-confidence word to edit.
VTT Generation
Processes items into cue blocks with timings.
Export
Download SRT/VTT files.
Results / Impact
- Improves caption accuracy.
- Speeds up post-transcription editing.
- Enhances accessibility.
Trade-offs / Limitations
- Requires AWS JSON format.
- No audio sync.
- Client-side only.
Next Steps
- Integrate with AWS API.
- Auto-suggestions for edits.
- Multi-language support.
Tech Summary
Vue.js, Bootstrap, JSON parsing, file download.
Summary Pattern Example
Problem: Edit AWS Transcribe output for captions.
Approach: Vue tool with confidence highlighting and VTT export.
Result: Efficient post-processing for accurate video captions.