aws-samples / aws-samples/amazon-transcribe-post-call-analytics
Use GenAI for languages not supported by Amazon Transcribe Call Analytics
- Dominant language
- Python
- Stars
- 109
- Forks
- 89
- PR merge metrics
- No merged PRs in 30d
Description
Amazon Transcribe Call Analytics currently supports a limited set of languages, creating a gap for many European languages including Romanian, Croatian, Czech, Dutch, etc. (https://docs.aws.amazon.com/transcribe/latest/dg/supported-languages.html)
I propose to create an alternative analytics pipeline for non-supported languages:
1. Use basic Amazon Transcribe for speech-to-text conversion
2. Implement GenAI-based analysis on the transcribed text to provide similar capabilities to Call Analytics
3. Maintain compatibility with existing Call Analytics structure
Would appreciate feedback on this.
Contributor guide
Research direction
Start by reviewing the existing Amazon Transcribe Call Analytics pipeline and its compatibility structure; the issue names no files, tests, or entry points. Define the alternative pipeline for unsupported languages, then verify that transcription and GenAI analysis provide comparable analytics while preserving the existing structure.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- aws, python
- Domain
- ai, backend
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100