27 Oct 2026

Multi-Cam, Speaker Detection, Reusable Assets, and Sequence Packages

Sequence 1.4.0 expands the transcript-led workflow to multiple cameras, reusable media, and editable project handoffs. It also brings on-device speaker detection, a People library, a more capable timeline, and new tools for working with compatible AI assistants.

Multi-Cam Editing

Bring camera, screen, and external-audio recordings together in a Multi-Cam folder, then edit a combined programme rather than working from one camera alone.
Review recording roles and associate camera angles with speakers before synchronization.
Use waveform Auto Sync to align matching recordings. Low-confidence or unsuccessful matches stay flagged for review, with timing adjustments available when needed.
Choose camera angles for transcript passages and refine the resulting cuts. Speaker associations can help build automatic angle decisions that you can review and edit.
Create alternate programme variants with their own edits and camera choices, and choose which editing variant each camera contributes.
Use shared external audio independently of the visible camera. Synchronization corrections remain separate from automatic offsets.
Compose picture-in-picture and split-screen views for supported camera and embedded-stream combinations.

Read the Multi-Cam guide.

On-Device Speaker Detection and People

Identify speakers locally with the optional Nemotron 3 speaker-detection model, separate from the transcription model.
Run speaker detection after transcription or enable automatic detection as part of the transcription workflow. You can rerun detection without retranscribing the recording.
Name and color speakers, correct assignments, and filter the transcript by speaker.
Manage speakers across projects within each Catalog through People, including names, avatars, and colors. Like Assets, People are Catalog-scoped.
Optionally save encrypted voice profiles in your local library to help recognize returning contributors across recordings within a Catalog. Review matches when recording conditions or microphones change.
Merge duplicate People entries or forget a person through the library controls.
Improved paragraph handling around speaker turns and unassigned words, while preserving manual paragraph breaks.

See speaker detection and People.

Reusable Assets and Timeline Layers

Organize images, video, and audio in catalog-scoped Assets and folders for reuse across edits.
Add an editing variant to Assets as a live link. Changes to that edit can flow through to the places where it is reused, without exporting another movie first.
Place assets into the timeline and as layers above the main picture, with independent timing and source in-points.
Adjust visual layer position and size in the player, and control layer stacking, volume, mute, and locking.
Move and trim placed media directly on the timeline while preserving its source offset.
Assets already used by project layers are protected from accidental library deletion.

Learn about layers and reusable assets.

Sequence Packages

Share an editable Sequence project as a .sqbundle, including its variants, transcript and speaker edits, clips, markers, keywords, trims, framing, and cleanup decisions.
Multi-Cam packages carry their related cameras, programme variants, angle decisions, synchronization, and shared audio. Linked assets and related editing variants travel with the handoff too.
Include verified copies of original source media, or omit them when the recipient already has access to shared storage.
Optionally encrypt the editing data and send the generated code separately. Included source media is not encrypted by this option.
Open a package by double-clicking, dropping it onto Sequence, or using Import. Relink missing sources during import when needed.
Preserve People associations across handoffs. Encrypted packages can also carry remembered-voice matching evidence when enabled, helping the recipient reconcile contributors with their own People library.
Package replacement keeps the previous bundle until the new export succeeds.

See Sequence Packages for handoff and encryption details.

Timeline, Trimming, and Transcript Editing

Added a resizable timeline with video filmstrips, audio waveforms, and separate lanes for placed media.
Trim the beginning and end independently for each variant without modifying the original recording.
Restore either trimmed end without losing the other. Restoring the complete recording clears the trim state.
Keep sentence and clip reordering consistent across the transcript, Clips, timeline, playback, and export.
Trimmed material remains available to restore in Clips. Markers outside the retained range remain listed as outside the trim rather than being deleted.
Silence rows and removal respect the retained source range, including pauses that overlap a trim boundary.
Improved the consistency of transcript refreshes, speaker assignments, and per-variant editorial choices.

Explore transcript editing and playback and trimming.

Multi-Track Recording Support

Inspect recordings containing multiple embedded video and audio streams, with separate choices for display, playback, and transcription.
Assign track roles and compose supported video streams as picture-in-picture or split screen.
Prepare audio mixes with per-component enable, mute, solo, gain, pan, channel mapping, and timing controls.
Preserve stream identities through relinking, and request review when a replacement cannot be matched confidently instead of silently selecting another microphone or camera.
Keep external-audio coverage tied to the actual recording, including sources that start late or finish before the picture.

See recording tracks and composition.

MCP and Command-Line Automation

Connect compatible AI clients through the Model Context Protocol, with setup and connection status in Settings → Automation.
Find projects, start a protected-copy editing session, review a rough cut, remove silence or filler words, and deliver the result through task-focused tools.
Editing sessions create a copy by default so an assistant can work without replacing the original cut. Saved changes remain when the session ends.
Added dedicated delivery tools for Final Cut Pro, DaVinci Resolve, finished video or audio, Sequence Packages, frames, linked Assets, transcripts, and subtitle files.
Expanded sq workflows for project discovery, transcript editing, cleanup analysis, automatic rough cuts, and background-job monitoring.
Agent activity is scoped to the affected projects, leaving other projects available while a command runs.
Automation remains optional and disabled by default. App-backed actions require Sequence to be open, signed in, activated, and unlocked.

Get started with the CLI or MCP setup.

Export Improvements

Extended finished-media export to follow supported Multi-Cam compositions, timeline layers, linked assets, and selected audio.
Improved Final Cut Pro handoffs for Multi-Cam structure, connected layers, native titles, camera source in-points, and picture/audio alignment.
Improved handling of audio boundaries and timing when camera and external-audio recordings have different start times or durations.
Preserve the effective profanity audio and redaction choices in Sequence Packages so a recipient’s different app defaults do not change the intended treatment.
Keep exported markers aligned with the retained edit, omitting markers in trimmed or removed material.

Editor handoff support varies by composition and destination. Review export notices and the guides for finished media, Final Cut Pro, and DaVinci Resolve.

Command Center and Workspace

Rebuilt the Command Center results list with native text input and virtualized rows for more consistent typing and navigation.
Fixed the dark Command Center turning light while typing or refreshing results.
Improved keyboard selection so a stationary pointer does not unexpectedly change the highlighted command.
Improved text selection and input-method composition while searching for actions.
Reduced initial-launch work by deferring optional background services until after the first startup stage.

Library Upgrades and Account Access

Added a recoverable upgrade path from the 1.3.8 library, preserving the original until the upgraded copy is ready.
Library upgrades now ask for access to the source folder when required and stop with actionable permission or Keychain errors instead of creating an empty replacement.
Added checks for unsupported newer libraries and simultaneous access before writing to the library.
Strengthened email sign-in with fresh, device-bound challenges and account-mismatch protection.
Improved sign-in feedback for connection problems, unavailable device keys, and resend limits. Email resend cooldowns survive closing and reopening the sign-in flow.

For upgrade and account guidance, see upgrading your library and account activation.