Directed Screen Films

Directed camera: Sit the recording in a device — MacBook Pro, MacBook Air, Studio Display, a phone, a tablet, a browser chrome, or an app window, posed on a stage instead of a flat rectangle

Directed camera: Drop the take into a real place — billboards, living-room televisions, theater screens, stadium boards, museum galleries, and an unbranded watch face, each with its own calibrated screen

Directed camera: Three film styles the camera follows on its own

Directed camera: Drop a marker with S at the playhead to say something happens here

Directed camera: B-roll at the cuts — named cutaways, camera moves and second-camera angles (over the shoulder, typing, above the deck, screen dolly) play over the head of the next clip

Directed camera: Light and finish — glare that tracks the turn, a sweep when a new device arrives, a contact shadow that breathes, background parallax, a light vignette, and a shutter smear on fast moves

Directed camera: Speed ramps through idle stretches, offered only on a take with no audio track — a recording that has microphone or system audio keeps a single rate so picture and sound stay together

Score: Score this take with a local bed that follows the film’s cuts, not a loop stapled underneath

Score: Compose with AI through the AI CLI you already connected — on your click, never automatically. The request is the shape of the film (cuts, mood), not the footage

Talking head, captions and the pointer: Record a camera track with the screen take and place it on the exported frame — rounded or circle, any corner, or full-frame

Talking head, captions and the pointer: Burned-in captions, word-level, transcribed on this machine

Talking head, captions and the pointer: Restyle the redrawn pointer — the macOS arrow and a pointing hand, plus character skins (cat, dog, chick, banana, gold trail, and more) that tap, peck or bat the click

Talking head, captions and the pointer: Capture keyboard shortcuts while you record — opt-in, Accessibility-gated, chords only (⌘ / ⌃ / ⌥). Bare typed text is never read

Studio editor: A real timeline with lanes for clips, the click zoom, clicks, masks, shortcut labels, markers and the score

Studio editor: Suggest trims for silence, filler words and idle stretches — suggestions only; applying one edits the cut and re-opens the review gate. Filler needs a transcript

Studio editor: Photographic wallpapers and plates — rooms the device can stand in, and backdrops the take sits in front of, next to the existing gradient grounds

Studio editor: An export dialog that asks for resolution, frame rate and quality immediately before the file is written

Studio editor: Shortcut and typing labels on a moment of the take or on a guide step — the shortcut you want the follower to press, or a description of what they type, never the characters typed into the recorded app

Screen recording setup: A preview of how the take will look before you press Start — the chosen screen, the redrawn pointer, the camera bubble and a sample shortcut pill

Screen recording setup: Sound, the take name and the frame rate sit in a right-hand rail, so the stage stays the picture: pick a source, see the frame, start

Screen recording setup: Continuous screen recording on Windows and Linux as well as macOS (macOS still prefers the native helper; video, microphone and pause on Mac still need macOS 15)

One recording, two editors: Jump between the video and the steps of the same screen recording, landing on the same moment

Annotation editor: Mark up a step’s screenshot with crop, arrows, text, shapes, counters, a pencil, a highlighter, a spotlight and an image overlay

Guide exports: SCORM 1.2 — a package an LMS can take; practice at the end of a chapter reports completion

Guide exports: PowerPoint — one slide per step, speaker notes from the step body

Library: Projects, folders and tags over the same library

Library: A grid of first frames so eight takes all called “Entire screen” are still distinguishable, plus the working list view

Library: An About card with version, a check for updates, and the support / email / releases links

The record picker leads with a drawing of what each source can and cannot see — browser (the only one that redacts before capture, and the only one a follower can replay), one-window stills, the auto screen recorder, and the screen recorder for video

Generation customisation is tabs, not stacked accordions — branding, languages, annotations, video and guide type stay glanceable and the Generate button no longer jumps

Describe the flow before Generate — a prompt with example chips, so the model is told who it is for without a second settings hunt

The capture bar is one compact row over the thing you are recording, instead of three stacked rows and a full-width note field

A theme toggle in the sidebar cycles Dark, Light and Match system

The camera will not outrun the screen — if a move would smear the UI past readable, amplitude comes down and the director panel says by how much

Changing how a take looks (style, intensity, grade, markers, b-roll, wallpaper, score, captions, pointer skin) leaves the review gate shut; changing which pixels ship (masks, trim, clip rates, speed ramps) asks you to watch it back

Project redaction rules now run in the browser recorder before anything leaves the page — the same words you saved still destroy the value at capture, not only later on the stored screenshot