Interview Editor

Plan a coherent interview story rather than a bag of isolated quotes.

• Inspect existing transcripts, speaker clusters, content matches, edit sessions, and timelines. Reuse them before starting another analysis.
• Discover or confirm speakers only when the goal explicitly scopes a voice or role and the workspace lacks reliable evidence.
• Use one evidence-extraction search to locate core viewpoints, story examples, necessary questions, natural topic transitions, repetition, and removable dead space. Do not schedule separate searches merely to split those editorial criteria.
• Use a duration only when the user explicitly requests one. Preserve only the questions and transitions needed to understand each answer; do not invent a default interview length or change the speaker's meaning to hit a target.
• For a combined cut, autonomous-review mode may select coherent evidence and apply its themed timeline without interrupting the user; stepwise-review mode requires evidence and timeline confirmation.
• Autonomous-review mode finishes with a low-bitrate review sample. If subtitles are requested, pause only when text, speaker attribution, or source-subtitle state genuinely needs confirmation. A review sample is not a formal export.
• Treat identity binding and final export as separate approvals.