High Level Scenarios

Common downstream uses:

• post-call summaries
• ticket enrichment
• searchable clip libraries
• internal review or handoff notes

Scenario 2: Batch S3 Archive Transcription

Use batch mode when call archives or media libraries already live in S3.

Flow:

1. Build a batch request with input prefix and output prefix.
2. Submit POST /aiservices/scribe/jobs.
3. Track state by webhook or polling.
4. Read /jobs/{jobId}/files for per-file success/failure.
5. Ingest outputs into search, analytics, or storage.

Common downstream uses:

• compliance and audit logging
• searchable webinar or podcast archives
• bulk transcript backfills
• QA scoring inputs

Scenario 3: Zoom Recording Export + Re-Transcription

Use when you must re-process Zoom-managed recordings with your own transcript settings.

Skill chain:

• zoom-rest-api to fetch/download recordings
• scribe to transcribe exported media

Typical reasons:

• you need your own retention/search pipeline
• you need different transcript settings than Zoom-managed defaults
• you want to enrich recordings with your own summarization or tagging flow

Scenario 4: Compliance / QA Processing

Use batch mode when transcripts must be generated offline for audits, QA scoring, or archival search.

Prefer:

• word_time_offsets=true when reviewers need precise excerpts
• channel_separation=true for stereo call recordings
• webhook + queue ingestion instead of synchronous polling for large volumes

Scenario 5: Customer Support Voice-to-Insights Pipeline

Use when support call recordings should feed operational analytics instead of stopping at raw transcript text.

Flow:

1. Ingest call recordings from storage or exported meeting assets.
2. Transcribe with scribe.
3. Store transcript plus speaker/timing metadata.
4. Run downstream sentiment, keyword, escalation, or QA logic in your own pipeline.

Guardrail:

• keep scribe focused on transcription
• do sentiment analysis, keyword detection, or scoring in downstream services after transcript generation

Scenario 6: Browser Microphone Incremental Transcript

Use when a web page should capture microphone audio and show transcript updates every few seconds without switching to RTMS.

Flow:

1. Browser captures microphone audio with MediaRecorder.
2. Browser flushes one chunk every 5 seconds.
3. Backend accepts each chunk as a normal fast-mode upload through the async wrapper.
4. Frontend polls by request ID and appends transcript chunks in order.

Guardrail:

• this is pseudo-streaming over repeated file uploads
• this is best kept as a lightweight demo or constrained fallback
• do not choose it first for a true live-transcription product
• if the requirement is truly live media stream ingestion or lower-latency continuous audio, route to rtms