Call and mail sample generator
Load config
Save config
Fill out KONE example
Configuration
Output folder
i
Working directory (fixed on server):
/app/.web_work
Number of calls / mail threads
i
Estimate: ~0 MB storage, ~$0.00 API cost.
GPT Model
i
Company Name
i
Company Info (Text)
i
Call center Info (Text)
i
Setup
i
Tip: use an LLM to adapt this JSON to your use case (domain, taxonomy, metadata rules, and allowed field values).
Copy setup to clipboard
Clear setup
Number of fictional call center employees
i
Audio Post-processing
i
Output Format
i
Sample Rate
i
Bitrate (e.g. 128k)
i
Speed Factor (e.g. 1.0)
i
Processing Steps (sequential pipeline)
i
1
Generate call metadata (scenarios, dimensions, clusters)
Artifacts: —
i
2
Generate call transcripts from metadata
Artifacts: —
i
3
Evaluate metadata quality and distributions
Artifacts: —
i
4
Evaluate transcript length vs. metadata targets
Artifacts: —
i
5
Synthesize call audio (text-to-speech + sidecar JSON)
Artifacts: —
i
6
Post-process audio (format, bitrate, sample rate, speed)
Artifacts: —
i
7
Generate customer–service email threads (JSON + EML)
Artifacts: —
i
Run Pipeline (~$6/100 calls)
Cancel Processing
Dummy mode (skip external APIs, no costs)
i
No job started yet.
Live Output
Live log
Working Directory
Refresh
Full server path:
/app/.web_work
Contents:
Output directory already exists
Cancel
Run anyway
Delete?
Cancel
Delete
OK
JSON Preview
Close
Audio Preview
Close
EML Preview
Close