After syncing your footage, click Start analysis to begin. Analysis uses credits from your plan balance. The process takes a few minutes, depending on footage duration, your internet connection, and hardware, and builds the foundation for all the tools that follow.
What the AI does
During analysis, Selects performs several key functions in parallel:
Transcribes all speech with word-level accuracy within 9ms
Identifies individual speakers through diarization with 97-98% accuracy
Detects topics and subtopics across your entire project
Analyzes visual content to detect scenes, cuts, and camera angles
Groups retakes of the same content and selects the best take to place on the timeline
The analysis progress screen
While the analysis runs, you see a progress screen that shows the current stage. This gives you visibility into what the system is doing and how much time remains.
What Selects generates after analysis
A Topic stringout: this is an uncut, unfiltered timeline with all your footage in chronological order and marked by topics.
Transcription: AI transcribes all your footage. Every word gets a time-stamp. You can use this transcription for text-based editing.
Learn more in about Text-Based Editing in [Refine Your Draft]
Speaker detection: AI identifies each speaker and automatically assigns them to the most relevant camera and audio tracks. This powers camera and audio switching on multi-cam timelines.
Learn more in []
Topics and Subtopics: AI uses the transcription to understand your content and organizes it into topics and subtopics. Use these to quickly understand what the footage is about, create drafts and more.
Learn more here What are Topics and Subtopics
Visual analysis: AI analyzes all media footage and adds a short visual summary to each clip. This is used when auto-adding B-roll to a timeline. It also helps when searching for a specific scene.
[Adding B-roll...]
[Search Your Footage: Text Search and Scene Search]
Retake grouping: AI detects retakes, groups them, and selects the best take with a reason for its choice. You can switch between takes at any time. Most useful for talking head and scripted content where the same line is recorded multiple times.
Credit usage
Analysis uses credits from your plan balance. Here is how credits are counted:
Video analysis: 1 credit per 10 seconds (6 credits/min, 360 credits/hr)
Image analysis: 1 credit per image
Audio analysis: 1 credit per minute (60 credits/hr)
Learn more about plans and credits []