Run OpenAI-Compatible STT in Daytona

[aqin236]

/claim #13

Summary

• add a long-form Daytona guide for running Sapat with a generic OpenAI-compatible STT endpoint
• add a glossary definition for OpenAI-compatible speech-to-text APIs
• add an author profile and an original SVG workflow diagram
• link the companion Sapat provider implementation: Add generic OpenAI-compatible STT provider nibzard/sapat#70

Validation

• git diff --check
• SVG parsed as XML with PowerShell
• guide word count: 1534
• scanned changed files for common token, private key, local path, Gmail, and noreply-email patterns
• companion Sapat PR validation: python -m pytest tests\providers\test_openai_compatible.py tests\test_registry.py -> 16 passed
• companion Sapat PR validation: python -m compileall sapat tests
• companion Sapat PR validation: python -m black --check sapat\providers\openai_compatible.py tests\providers\test_openai_compatible.py

Note: I did not run markdownlint locally because this checkout does not have node_modules, and I avoided installing new dependencies in the workspace.