Start a recording session: launches a themed headless browser and begins capturing video. Returns the session config and the output directory. One session at a time — studio_finish or studio_cancel ends it. Set theme to the PRODUCT'S brand (bg = its app background — wrong bg causes visible flashes on page loads); use storage_state_path (a Playwright storageState JSON) to record logged-in areas without sharing credentials; list chat widgets / dev overlays / consent banners in hide_selectors so they never appear on tape.

Navigate to a URL behind a designed transition card (eyebrow + big title on the theme background) — the viewer never sees a loading flash or half-hydrated page. Use for the opening shot AND every surface change; give each a short title so the cut reads as a chapter. Waits for real content (network settle + optional wait_for selector + skeleton loaders cleared) before lifting the card.

Glide the visible cursor to a target and click it, slow enough to read as a deliberate action on camera. Use for every button / link / menu interaction in the walkthrough. Precondition: the target must be on screen — studio_goto or studio_scroll to it first if it isn't. Prefer a stable CSS selector; fall back to visible text. After clicking it waits for the page to settle and re-applies the theme overlays, so the next beat is on a hydrated page. Returns 'clicked ', or errors 'target not visible' — then take a studio_screenshot and pick a better target.

Glide to an input and type text at a human pace (keystroke by keystroke — it reads as real usage, not a paste). Use realistic demo content; whatever you type is on tape.

Cubic-eased scroll to an absolute Y offset or until a target element is in view — never a jump cut. Use between story beats to bring the next area on camera. Side effect: changes only the page scroll position (no click, no navigation). Pass exactly ONE of to_y or to_target. Returns 'scrolled' on completion. To reveal an off-screen element before studio_click or studio_highlight, scroll to it first.

Show (or replace) the bottom-left storyline caption — an eyebrow kicker plus one sentence that narrates the current beat. Keep it under ~12 words and leave it up 2–4s (studio_hold) so it reads. Pass clear=true to fade it out before a scene change. The caption_text is ALSO recorded as the voice-over script: if you call studio_finish with voiceover:true it's spoken aloud (locally) timed to this moment — write captions as speakable sentences. Set no_voice:true for a caption you want on screen but NOT narrated.

Record a line of voice-over WITHOUT putting a caption on screen — narration spoken over the action. The line is synthesized locally at studio_finish (needs voiceover:true). Holds the shot afterward so there's footage under the line; set hold_ms to roughly the length of the sentence when spoken (~400ms/word).

Draw an accent ring around a target (CSS selector or visible text), optionally with a label card beside it — the 'look here' callout. Side effects: scrolls the target into view, then overlays a ring (and card) in the theme accent; rings stack until cleared. Returns 'highlighted ' when the element is found, or errors 'target not found: ' when it isn't — retry with a studio_screenshot to pick a better target. Always call with clear=true before you navigate, zoom, or ring a different element, so stale rings don't linger on the next shot.

Camera punch-in: smoothly zoom the page toward a target (CSS selector or visible text) — the money shot for a metric, button, or result. Captions stay screen-fixed; the content scales under them. ALWAYS studio_zoom with reset=true before navigating or showing the end card, or the next scene inherits the zoom.

Hold the shot for a moment — glides to a target (optional) and keeps micro-drift so the recorder keeps emitting frames (a dead-static hold gets its tail frames dropped). Use after captions, highlights, and zooms: 2000–3000ms is a good beat.

Show the closing card — a centred glass card over the (still visible) product: title, optional subtitle and URL. The classic outro: what they saw + where to get it. Hold ~2.5s after this, then studio_finish.

Take a high-DPI (2x) screenshot of the current frame and save it as a PNG in the output directory. Works any time during a session — element-only via selector, or full_page for the whole scroll height. Returns the file path. Clear captions/rings first unless you want them in the shot.

Stop recording and produce the final files: closes the browser, strips residual dark frames (ffmpeg blackdetect → trim), transcodes to a share-ready MP4 (H.264, faststart), then — optionally — narrates the caption lines with a LOCAL voice (no cloud/keys), lays down a music bed, writes/burns subtitles, and exports vertical/square social crops, plus a GIF. Voice-over timing is auto-corrected for the frames the deblack step removed. Returns every produced file with size + duration. Without ffmpeg you still get the raw WebM.

Abandon the current recording session: closes the browser and discards the tape. Nothing is written except screenshots already saved. Safe to call when no session is active (no-op).