Hermes contribution
Speech transcription for Telegram video notes
This open PR proposes transcribing Telegram video notes through the existing speech-to-text (STT) path when STT is enabled. The video remains available for visual analysis. Ordinary videos are not sent for automatic transcription, and an already-transcribed pending message skips a second STT pass. This follows the video routing merged separately in PR #147572 rather than adding that routing again. PR #147584 is not merged.
This Hermes contribution comes from the source PR. See the source PR for status and discussion; release inclusion is checked separately.
GitHub data verified:
PR status
Open — proposal
PR inclusion in a release is unverified
Changed files (6)
gateway/run.pygateway/run_inbound.pygateway/run_media_predicates.pyplugins/platforms/telegram/adapter.pytests/gateway/test_telegram_documents.pytests/gateway/test_telegram_video_note_stt.py
Topics for this object
Help with this task
AI consultation and workflow review
Diagnosis of the workflow, bottlenecks, risks, and ROI: what is worth automating, where an agent helps, and where manual control is safer.
Discuss the taskAI agent setup for a specific business task
An agent for a concrete process: data sources, rules, quality checks, safe autonomy boundaries, and ongoing support.
Discuss the taskTelegram bots / backend / API / automation
The implementation layer after diagnosis: bots, APIs, integrations, monitoring, internal panels, and self-hosted Node.js tools.
Discuss the task