Run Sapat with DeepInfra in Daytona

[anamanlab]

/claim #13

Summary

• Add a long-form Daytona guide for running Sapat with DeepInfra in a repeatable transcript intake workflow.
• Add a Transcript Intake Pipeline definition.
• Add first-time author profile for Arthur Porfirio de Souza.
• Add original SVG assets for the workflow diagram and author image.

Companion implementation

• Sapat provider PR: Add DeepInfra transcription provider nibzard/sapat#20
• The companion PR adds --api deepinfra, documents DeepInfra environment variables, supports Sapat's existing --correct flow through DeepInfra's OpenAI-compatible chat endpoint, and includes mocked unittest coverage for the transcription payload and corrected-transcript chat call.

Validation

Daytona content PR:

• npx --yes markdownlint-cli guides\\20260520_guide_deepinfra_sapat_transcript_intake.md definitions\\20260520_definition_transcript_intake_pipeline.md authors\\arthur-de-souza.md
• PowerShell XML parse check for both SVG assets
• git diff --check
• Get-Content guides\\20260520_guide_deepinfra_sapat_transcript_intake.md | Measure-Object -Word -Line (1,523 words)

Companion Sapat PR:

• PYTHONPATH=src py -3 -m unittest discover -s tests (2 tests)
• PYTHONPATH=src py -3 -m py_compile src/sapat/transcription/deepinfra.py src/sapat/script.py
• PYTHONPATH=src py -3 -m sapat.script --help
• git diff --check

Sources checked

• Sapat README/source
• DeepInfra audio transcription API docs
• DeepInfra Whisper speech recognition docs
• DeepInfra OpenAI-compatible chat docs
• Daytona contribution guidelines

Transparency: AI-assisted with Codex; content was source-checked and locally validated before submission.