Azure AI Speech Transcription with Sapat

[goodguypeci-design]

Closes #13.

Sapat already supports Azure OpenAI, but that route still expects a Whisper deployment. I added a separate Azure AI Speech provider in Sapat PR #80, then wrote this guide around the current Daytona sandbox workflow: creating the environment, configuring the Speech resource, uploading a short sample, checking the transcript, and diagnosing the failures a reader is most likely to hit.

The submission also adds the matching speech-to-text definition, my first-contributor profile, and an original labelled workflow illustration. The guide is 1,624 words. Targeted markdownlint and git diff --check pass, the illustration was rendered and inspected at 1600×900, and the DCO check is green. No local paths, credentials, recordings, generated transcripts, or payout details are included.

/claim #13

[goodguypeci-design]

@algora-pbc @zcesur Could you take a look at the claim sync for this PR? /claim #13 is already in the body, but no claim was created. The Algora tracker on issue #13 has not updated since June 19, and another contributor reported the same problem for PR #316 in the bounty chat.

My attempt is #13 (comment) and my Algora profile is https://console.algora.io/@/goodguypeci-design. Could you resync the claim, or confirm whether this bounty is no longer payable? Thanks.