Offline Transcription with Sapat and Daytona

[telemarkdigital-publisher]

Summary

• add a long-form Daytona guide for running local faster-whisper transcription with Sapat
• add a new Daytona definition for local speech-to-text
• add a minimal Telemark Digital Publisher author profile for this first content contribution
• add an original SVG workflow diagram asset

Bounty / issue

Closes #13
/claim #13

Companion implementation

The guide is grounded in the companion Sapat provider PR: nibzard/sapat#76

Validation

• Targeted markdownlint passed for the guide, definition, and author profile using workspace-local npm cache
• Guide word count: 1,766 words
• SVG XML parse passed
• git diff --check passed
• Local secret/private-path scan found no local filesystem paths, tokens, or private account references
• Companion Sapat validation: pytest full suite passed with 182 tests

No paid APIs, hosted transcription calls, private media, or secret values were used. AI-assisted with OpenAI Codex and manually reviewed before submission.