{"version":"2.1.202","anchor":"audio-file-transcription-via-mention","canonical_anchor":"audio-file-transcription-via-mention","heading":"Audio File Transcription via @-mention","tier":null,"area":null,"scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.202\/e\/audio-file-transcription-via-mention","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.202","markdown":"### Audio File Transcription via @-mention\n\n**What**\n\nUsers can now @-mention audio files in their prompts. Claude Code sends the audio to Anthropic's speech-to-text service and injects a structured transcript into the conversation before Claude sees it.\n\n**Details**\n\n- The transcript is wrapped in `<audio-transcript filename=\"...\">` XML with optional `duration=\"M:SS\"` metadata\n- If transcription fails, an error variant is injected instead of silently dropping the attachment\n- The reference panel regex that handles pasted\/attached items now recognizes `[Audio #N]` placeholders alongside existing `[Image #N]` and `[Pasted text #N]` entries\n- Governed by a new HIPAA-compatible permission: `allow_voice_file_transcription`\n- Audio attachments are now passed through in subagent message filtering alongside image attachments\n\n**Evidence**\n\nAudio transcription handler (search for `\"audio-transcript\"` and `\"The user @-mentioned the audio file\"`)"}