Episodes
Episode management -- properties, media, outline, script, analysis, dialogue, storyboard, shot list, sequencer, progress, and guide.
Episodes are the primary unit of output -- a single installment in your series. The Episodes page lists all episodes in a project.
Click New episode and supply a display name; a short ID (lowercase, digits, hyphens) is suggested from it and you can edit it before confirming. Under the ID, one row of four short fields numbers the episode -- Season, Episode, Suffix and Part (see Episode numbering below). They start on the next free number, so the first episode is season 1, episode 1, and later ones continue the latest season. The new episode opens on the same page as opening any episode does -- its Properties, unless you came from another episode page. Asking the assistant to create one takes you there too. The list's one column is Title; an episode that has numbers also shows its episode ID (for example S01E03) beside the title, or under it on a card.
| Action | Notes |
|---|---|
| Select | Opens the episode, on its Properties view |
| Rename | On the row's "..." menu -- opens a dialog for the display name |
| Delete | On the same menu, with a confirmation |
| Star / Unstar | The row's ... menu. A starred item shows a yellow star beside its name |
List and grid
The control at the right of the controls row switches between a grid of cards and the list. Cards are what you get to begin with, and whichever you pick is remembered for next time.
A card shows the episode's picture with its title underneath. The picture is the last image generated anywhere inside the episode: on the episode's own Media page, on any of its scenes, or on any of its shots. Generate a newer one and the card moves to it. Delete the picture that is showing and the card does not fall back to an older image of the episode's. An episode with no picture of its own shows a plain, blank card until its first image is generated.
Cards carry the same star and "..." menu as the rows, in the same places as a media gallery card: star on the left, menu on the right.
Project review
Each episode reviews the project instead of repeating it. The decisions you made once for the whole show -- its look, visual styles and recipe choices -- are signed off on the project, and no episode asks you for them again. On the episode's Properties view, below Core properties, a small Project review card has one row: a checkbox, Review the project properties, and a Review button that opens Project Properties.
Ticking the box says you looked at the project's properties for this episode. It does not approve anything on the project, and it does not tick any other episode. A choice is always changed on the project, never for one episode, so every episode sees the same value. The review is a suggestion: it never stops you from going on. The guide bar offers the same step -- Review the project properties with Go, and I'm done once you are on Project Properties.
Major characters are reviewed where you choose them. After you select an episode's major characters, the guide offers Review each of them: it opens that character's Properties, and I'm done there ticks the review for this episode. Defining a character is still done once, on the project.
Properties
Selecting an episode opens its Properties view. Episode properties are grouped into cards, in the order the pipeline asks for them: Core properties, Story drivers, Story structure lock, Script review, Music & sound, Appearance, and Final review. Each card header can approve its whole set at once, and offers a Review with AI pass.
Core properties -- Episode name, the four numbering properties (Season number, Episode number, Episode suffix, Part number), Description, and Notes. The description is what identifies the episode in lists; notes are yours alone and are never read by the AI.
Episode numbering
Four Core properties give an episode its ID, written SnnEnn with an optional suffix letter and part: S01E01, S01E01a, S01E01aP01, S01E01P02. Numbers under 10 get a leading zero.
| Property | What it takes |
|---|---|
| Season number | A whole number, 1 or more -- season 1, season 2. (The whole show is the series; this is one season of it.) |
| Episode number | A whole number, 1 or more, within the season |
| Episode suffix | Empty for the full episode, or one lowercase letter for another piece of media made for the same episode -- a promo video, say |
| Part number | Empty, or a whole number, 1 or more, when an episode is released in parts |
Anything else -- letters in a number, punctuation, zero, two letters in the suffix -- is refused with a message rather than quietly changed. All four are optional.
The ID is a label, not a key. Nothing in the project looks an episode up by it, so two episodes may share the same numbers (a promo and the full episode do, by design), and changing a number never breaks a link.
An episode with no season number or no episode number has no ID and is not on the Timeline. Episodes made before numbering existed start that way: give them numbers on their Core properties, or ask the assistant to.
Story drivers -- the properties relevant to the writing and production of this specific episode:
| Property | Purpose |
|---|---|
| Episode objective | What this episode accomplishes in the arc |
| Core conflict | The central dramatic tension |
| Relationship dynamics | How character relationships evolve |
| Arc integration | How this episode fits into ongoing story arcs |
| Continuity notes | What carries over from the previous episode |
| Setup for next episode | Seeds planted for upcoming episodes |
| Premise | The episode's core premise |
Appearance -- the two picture sizes this episode is made at, each picked from the same list of sizes (every entry shows its name and its width and height):
| Property | Purpose |
|---|---|
| Rough clip size | What this episode's clips are first made at: small and cheap, so a shot can be tried many times. Recipe runs and the image layering step in this episode start at this size. |
| Finished size | What this episode is delivered at: the size the Sequencer shows and exports, and the size The Finish upscales to. |
A new episode copies both values from the project's Properties page (Appearance) when it is created, and keeps its own from then on. Changing the project's values later does not change episodes that already exist, and changing one episode's does not change another's -- so a project can hold vertical shorts beside full 16:9 episodes. An episode that has no value of its own shows the project's. The pipeline asks for them in the Episode Appearance step, at the start of The Visuals, before any boards or clips are made.
The later cards are review gates rather than authoring: Story structure lock records what is still blocking the structure; Script review holds your own notes on the first draft (beat coverage, clarity, motivation, dialogue, pacing, and the revision plan); Music & sound and Final review hold the equivalent notes for the audio and the finished cut.
These fields are used by the AI Assistant for context and by the Episode Plan's acceptance criteria.
Outline
The Outline view provides a high-level structural view of the episode -- scenes, beats, and the overall shape of the story. Use it to get an overview before diving into the script or individual scenes. It is also where the Story Consistency Check runs.
Script
The Script view is the main writing surface for the episode -- the screenplay, written in Fountain format. See Script for the full reference.
Analysis
The Analysis view runs an AI pass over the script and breaks it into the characters, locations, and props it mentions, so you can resolve each against your entity library. It also lists the music each scene calls for and the specific sound effects that carry the story (silence is a valid answer, so a scene may list no music); each of those needs has a tick box. Source the audio however you like -- reuse a piece from the project, generate it with an AI audio recipe, or upload it -- then tick the need off as done; the plan's Music and Sound effects criteria are met once every need is ticked (with no needs they owe nothing). The plan also fills a scene's music cue from its music need when the cue is empty. See Script for how analysis fits into the writing flow.
The Analysis view has two actions side by side. Analyze screenplay reads the script forwards, into the report below. Fill from screenplay (an AI button, marked with the sparkle) reads it backwards: it reads the screenplay back into the earlier steps, so that later work has the details it needs to spot a contradiction. One click does the whole job; behind it the app makes several AI calls, one per part. It proposes what the script already answers for:
- the project -- every property you can type or choose on the project's properties page that the script speaks to: description, identity (logline, theme, story type, series format, genre, tone and aesthetic, audience, episode structure, music), world building, appearance notes, and so on. A proposal here changes the project's own value, the same one every episode reviews.
- the episode -- every property of the episode the script speaks to (the story drivers, the premise, the music vision and so on). Your own review notes on the draft are not filled.
- each character, location and prop -- every property of its page the script speaks to: a character's role, goals, backstory, personality, build, hair and wardrobe; a location's layout, style, colours and mood; a prop's form and materials. No images are made, no voice is picked, and nothing is measured.
It only proposes what is true of the character, place or thing everywhere in the story. A condition that holds in one scene -- a character who is dirty in scene 3, a room on fire in scene 7, a cup that breaks in scene 4 -- is never proposed, and is never added to a value you already wrote. A property the script says nothing about gets no proposal, and a choice you have already made is not asked again.
A field that is still empty is filled. A field that already has text is proposed again as your text with what the script adds put after it: your words are kept exactly, and a field the script has nothing new for gets no proposal. Run it again after you edit the screenplay and it picks up the new facts. Before it runs, it says which parts it will propose and how many AI calls that is, and asks you to confirm. Afterwards it reports how many proposals it left.
The proposals are reviewed in one list on the Analysis view, grouped Project, Episode, Characters, Locations and Props, with a heading for each project part, character, location and prop. Each row shows the property, what is there now and what is proposed, with Accept and Reject. You review each proposal on its own: there is no "accept all". Nothing is filled in by the run: a value is written only when you accept it, and signing a step off is still your decision. The list is still there if you leave and come back, and a proposal you accept on a character's, location's, prop's or the project's own page leaves the list too.
After the analysis, the guide offers Fill in the earlier steps from the screenplay as a step of its own. It is ready whether you did the earlier steps, skipped them or never opened them; Go opens the Analysis view, where the two actions are at the top. Sign it off once you have looked at the proposals. If you edit the screenplay afterwards, the guide offers it again.
This is the only place the screenplay is read back into properties: no properties page or guide entry has a fill button of its own. Scenes and shots are filled forwards, from the screenplay's own scenes, not by this button.
Dialogue
The Dialogue view lists every line of dialogue in the episode and turns it into voice -- generate the whole episode's spoken track in one press, then work line by line through takes, recordings, and voice changers. See Dialogue Audio for the full reference.
Storyboard
The Storyboard view shows the episode's shots as a grid of 16:9 frames, in order, grouped under their scene headings -- the "see all the shots in order" overview. Each cell shows the shot's clip if it has one, otherwise its storyboard still, otherwise a "no media" placeholder, so the holes are obvious at a glance. Double-click a card (or choose Open shot from its menu) to open the shot; click a scene heading to open the scene. Shots can be dragged into a different order within their scene, and a round + appears in the gap between cards to insert a new shot there. Hide Shelved, at the right of the heading, leaves shelved shots and shelved scenes out while it is on; the heading still counts them. It is the same switch as the Shot List's, so turning it on in one view turns it on in the other.
Only pictures appear here. A shot collects more than pictures -- rendered dialogue gives it one audio take per line -- and this view is about seeing the sequence, so a shot whose chosen clip is audio (or a prompt, or a pose file) shows its storyboard still instead, and a shot with no picture at all shows the "no media" placeholder. The choice itself is untouched; only this view declines to draw it. The Scene images rail down the left, which you drag stills from onto shots, lists pictures for the same reason.
Each card carries a ... menu in its top corner -- the same button the media gallery's cards use. Besides shelving, duplicating and deleting the shot, it offers Remove as storyboard and Remove as rough clip when that choice is set, so a wrong frame can be taken back from the place where you can see it is wrong. The card falls back to the shot's other choice, or to the placeholder.
Shot List
The Shot List view shows every scene and its shots for the episode in one place -- a quick way to see the shape of the episode's coverage and jump to any shot. It presents them three ways, chosen with the List, Detailed and Reel view buttons at the right of the Shot List heading. Reel (the shot icon) is for finding a shot by what it looks and sounds like. Shots stay under their scene's heading, but instead of rows each scene shows its shots' media large, left to right and wrapping into rows, filling the window width (a wider window fits more per row). Each frame is the picture the sequencer shows for that shot: its clip, otherwise its image. A clip shows its first frame with a play button in the middle, the same as in a media gallery; tap it to play the clip right there, with sound, and tap again to pause -- only one plays at a time. Tapping plays the shot's full sound -- dialogue, music and effects, with their fades and levels, just as the sequencer plays that stretch of the episode. A shot with no picture but with dialogue shows an audio tile with its first line, which plays when tapped; a shot with nothing shows a "no media" placeholder. Tapping a picture only ever plays it -- it never leaves the page. Under each frame one line gives the shot's number and name ("1.12 Eyes power on") on the left and its length in seconds on the right; tap the name (it underlines when you point at it) to open the shot. A typical wide window fits six frames per row. Hide Shelved, the button beside the view buttons, leaves shelved shots and shelved scenes out of all three views while it is on; the heading still counts them. The rest of this section describes List and Detailed. Both show the same scenes with the same shots under them, as the same rows; Detailed simply puts more into each row, so the switch never loses your place and nothing moves sideways when you flip it.
List is the pacing view: each shot's allocated duration in a lined-up column, its start offset within the episode, its media state, and a total running time at the foot. It deliberately shows no pictures -- it reads as a printable document.
Detailed is the same list with the answer to a different question in it: are the inputs complete and correct before you generate? Each row keeps its start time, number, name and duration exactly where List has them, and gains a thumbnail column on the left, plus the shot's inputs underneath the name -- location, cast (with mood), and props, each badged from scene when it is inherited or override when the shot sets its own, the backdrop details inherited from the scene (lighting, palette, atmosphere, music, ambient), and the shot's dialogue lines. Nothing is edited in place: clicking a field, the name or the thumbnail opens that shot.
Shots reorder and insert the same way in both presentations: drag the handle at the left of a row to move a shot within its scene, and use the + beside it to insert a new one below (the + on a scene heading adds one at the top of that scene).
Sequencer
The Episode Sequencer is where the episode is cut. It shows the whole episode in one view -- the scenes as a strip, the shots as a strip inside them, and below those the tracks holding dialogue, music and effects -- and you can both watch it and change it here.
Watching it
Play, pause and step a frame at a time from the transport in the middle of the toolbar. Zoom to jumps the view to a whole episode, a scene or a single shot, and Set Range limits what plays and what exports.
The speaker button opens the sound panel, holding mute and the volume slider together. The button itself keeps showing the state after you close the panel -- crossed out when muted, and otherwise roughly how loud it is.
The size dropdown in the toolbar shows and changes this episode's Finished size (the same value as on the Episode Properties view); another episode keeps its own.
A shot you have not picked an image for yet still shows you something. If the project already has a starred picture of a character who speaks in the shot, or of its location, that stands in for the missing frame. If there is nothing to borrow -- which is normal on a new project -- the shot plays as a hatched panel naming the shot and saying it has no image yet, so you can hear the cut and read where you are instead of watching black. That panel is a note to you, not a frame -- it is never included in an exported video, where the shot stays blank.
Choosing what a click does
Next to the volume is a row of four tools. They change what a plain click means; everything you drag behaves the same whichever tool is chosen.
| Tool | A click does this |
|---|---|
| Select | Selects what you click. Double-click, or the (i) at the right end of a shot or clip bar, opens its properties. This is the everyday tool. |
| Split | Cuts the shot -- or the clip -- you click, at the point you click. The two halves add up to the original length, so nothing after them moves. |
| Insert | Press and drag to the right to open a gap. The green line follows your cursor, and the sequencer moves as you drag so you can see the result before you let go. |
| Remove | Press and drag to the left to close time up. The sequencer moves as you drag, so you are watching the result; let go and it holds there with a tick and a cross beside it. |
The cursor changes with the tool and a line under the sequencer reminds you what the click will do, so you can always tell which one is active.
Insert and Remove are the pair you will reach for most once the music is in: music usually runs across many shots, and these are how you bring the shot lengths into line with it. They are the same gesture in opposite directions -- Insert pulls time apart to the right, Remove closes it up to the left -- so dragging the wrong way in either tool simply does nothing.
Inserting time holds the shot you pressed in for longer, and the scene grows to match. The sound moves with it: anything in that scene placed at or after where you pressed slides along by the same amount, and anything before it stays exactly where it is -- so nothing you have already lined up drifts out of sync. No cut moves, so nothing you have already cut gets re-cut.
Removing time shows you the answer, not the cost. As you drag left, the shots and the sound after the point slide left with your cursor, already sitting where the removal will leave them -- so you stop when the picture looks right rather than counting seconds. The playback range moves with them too.
Because removing is the destructive half and there is no undo, it is deliberately two steps: letting go holds that picture instead of committing it, and a tick and a cross appear at the seam. The tick makes it real. The cross puts everything back, and so does Escape, choosing another tool, or simply starting a new drag. Nothing at all is written until you press the tick. (Delete and the wastebasket button do the same thing as the tick, if your hand is already on the keyboard.)
When you do commit, the span stops existing and everything after it closes up:
- Shots that reach into the span from either side are shortened. A shot lying right across it keeps playing and simply loses that much time. A shot wholly inside the span is hidden, not deleted -- it keeps its length and its place, and you can bring it back from the Shot List.
- Sound that reaches into the span is trimmed at that edge. A music bed lying right across the span is shortened from its end and never cut in the middle, so it keeps playing without a hole in it -- and you can drag its end back out if you took too much. Anything wholly inside the span is unplaced; the file itself stays in the project.
A span has to sit inside a single scene. To take time out of two scenes, do them one at a time.
Changing the cut
- Move a cut by dragging the diamond that sits on it. The shot before grows, the shot after shrinks, and the scene stays the same length -- so every sound in the scene keeps the position you gave it.
- Remove a cut by clicking the diamond and pressing Delete, or the wastebasket button in the toolbar. The two shots merge into one. Nothing is thrown away: the shot that was absorbed keeps its record and anything you rendered for it.
- Lengthen the last shot in a scene by dragging the grip on the right-hand end of the scene. That one grows the scene.
Muting a clip, and sliding a clip inside its shot
Every video clip that has its own sound, and every dialogue, music and effects clip, carries a small speaker on its bar. One tap mutes that clip and another brings it back; a muted clip shows a struck speaker and a dimmer waveform. Where a shot has both a talking clip and dialogue clips for the same lines, only one of them is heard to begin with: the clip's own sound, and the dialogue clips show muted. Tap a speaker to choose the other. The preview, the Shot List reel and the exported video all follow what the speakers show. The track's own mute and solo still work on top.
Select a video clip on the Video row to see how much footage the shot is leaving out. A darker shadow shows the whole clip running past the shot's cuts (over the neighbouring shots when it reaches into them), labelled with the amount beyond each edge ("+1.2 s", "+2.7 s"). Nothing shows when the clip has no spare footage, and a still image, a looping clip or a clip of unknown length never does. Drag inside the selected clip to slide it: a different part of the clip plays in the same shot, while the shot, its cuts and its length stay put. The drag stops where the footage ends, and the preview shows the frame at the playhead for the part you have slid to as you drag (if the playhead is in another shot, it shows the first or last frame of the clip you are sliding). To change how long the shot is, move its cut.
The playhead moves only when you click or drag in the time ruler along the top. Clicking a clip or a track selects it and leaves the playhead where it is, and clicking the ruler keeps your selection, so you can set the playhead and then slide the clip you have selected.
Bringing the script's dialogue in
Every line of the episode's dialogue is laid onto a dialogue track in script order for you, as soon as you open the Sequencer. There is nothing to press. You do not need shots for this: a scene that has nothing but a script and rendered takes can be played straight through, which is how you judge pacing and delivery by ear before any picture exists.
It says what it did -- "3 lines to add, 1 no longer in the script" -- rather than changing the sequencer silently. Dialogue you have already placed keeps the timing you gave it: a clip already playing a line's audio is recognised and left where it is, never duplicated. After a script change, only what is new is brought across.
What it does, line by line:
- A line that has a take and is not on the sequencer yet is added. On an empty scene the lines are spaced out in script order. On a scene you have already worked on, a new line lands after the line before it, and everything later in that scene slides along to make room -- so the pacing you settled stays settled.
- A line you have re-rendered takes its new take's length. A longer take pushes what follows along; a shorter one leaves a gap.
- A line the script has moved is laid down again in its new place, and the clip it used to have is left where it was, marked.
- A line that has left the script has its clip marked, not removed.
Nothing is ever deleted and nothing is ever pulled backwards -- time is only made. Clearing the gap a marked clip leaves is an edit only you should make, because sometimes the silence is the point; use the Remove tool when you want it gone.
Lines with no take yet stay off the sequencer, and the dialog says how many there are. Generate them on the Dialogue view first.
If the episode has no tracks yet, the button creates a Dialogue, Music and SFX track to place on.
The track column
Down the left of the sequencer is a fixed column naming every row -- the scenes, the shots, the picture, and each sound track under them. It stays put as you zoom and scroll, because a track's name is a fact about the track and not about any moment in it.
Every row carries an icon at the left edge, so the list can be read down the column without reading the names: a signpost for Scenes, the shot grid for Shots, a camera for Video, a speaker for every sound track, a face for a mood track, and on the Playback Range row the same icon the Set Range button in the toolbar carries.
Double-click a track's name to open it. That is where you rename it, set its level, mute or solo it, and choose whether it ducks under dialogue. The (i) at the right of the name opens the same thing.
A track's name decides what lands on it. A track whose name starts with "Dialogue" is where generated lines go, and it is what other tracks duck under; one starting with "Music" is where generated music cues go; "SFX" marks effects. Rename a track and you change what is placed there automatically, so the dialog tells you what the name you have typed currently declares.
+ New track, under the last one, adds a track at the bottom. It asks which kind you want, as a pair of choices:
- Add a new sound track, which holds media. It arrives with a sequential name that means nothing -- rename it to give it a purpose.
- Add a new mood track, which holds text saying what one of your characters is feeling. Picking it lets you choose the character; you cannot add one without choosing. See The mood curve below.
Nothing is added until you press Add track, so you can read both choices before committing to either, and Cancel leaves the sequencer untouched.
Working with sound
Drag a clip along its track to change when it plays, or up and down to move it to another track. Drag the lower part of either end to trim it. What snaps as you drag is the moment inside the sound you are lining up -- a door slam, a first syllable -- not the start of the file.
Dragging an end inwards trims material off; dragging it back outwards puts that material back, and stops once there is none left. So a clip can never be stretched past the end of its own file. Stills are the exception -- a single frame holds for as long as you ask it to -- as are clips set to loop, which simply repeat to fill whatever length you give them.
Fades live along the top edge. Each end of a clip has a small triangular handle there: with no fade set it shows as a cut-off corner, and dragging it inwards fades the clip in or out. The sloped wedge that appears is the fade itself -- the slope is the volume ramp. Once a fade is set the handle sits at the point where the fade ends, so you can always see how long it is and take hold of it again to change it. Double-click the clip to read both fades as numbers.
The anchor is the small triangle standing on a clip's bottom edge. It marks the moment inside the sound that you are lining up with the picture -- the gunshot, not the second of room tone before it. That moment is what holds the clip in place and what snaps when you drag, which is why a clip with a long run-up sits further left than you might expect. A clip whose anchor is right at its start shows no triangle, because its left edge already marks it.
Double-click empty space on a track to add something at that moment (on a touch screen, just tap the empty space); the track's name decides what it offers you. Double-clicking past the end of the episode works too: the last shot is lengthened until that moment exists, and the sound is placed there. Nothing before it moves, so nothing you have already lined up drifts.
A sound placed this way arrives at its own full length, wherever in the scene you drop it -- a short sound dropped a moment before the scene ends is not cut down to fit, the scene grows to hold it. Two things can still shorten it on arrival: the next clip on the same track, because a track plays one clip at a time; and a file long enough to resize the episode on its own, which arrives as its first ten seconds rather than dragging everything after it out of shape. This only ever shortens the clip, never the file -- drag its end outwards to get the rest back. The line under the sequencer says when it has been shortened and what stopped it.
Delete removes a selected clip straight away. Only the placement goes -- the file stays in the project, so putting it back is a double-click rather than a re-render.
The mood curve
The Sequencer can say what is heard at a moment and what is seen at it. A mood track is how it says what a character is feeling -- the thing an image or video model most needs told, and the thing you most want to shape over time. A line is rarely just "angry": it starts annoyed, turns, and is angry by the end of it.
Add one from + New track: choose Add a new mood track, pick the character, and press Add track. It arrives already covering the episode: one clip per scene, spanning the whole of that scene, saying neutral. A mood track is never empty and never has holes in it -- there is a feeling in force at every instant, which is the whole claim the track makes.
One mood track per character. A character who already has one on this episode is not offered again -- two tracks would be two answers to a question that has one, and nothing reading the curve could say which was true. If everybody already has a track the list says so; delete a character's track to start theirs over.
The row is named for the character, and nothing else -- you can only ever see one person on a mood track, so a prefix would be width spent saying what the icon already says. It follows the character: rename them and the row renames itself. You cannot rename a mood track directly, for exactly that reason.
A mood clip holds text, not sound. It is never heard, never loaded and never exported; no waveform is drawn on it and it carries none of the controls a sound has -- no fades, no level, no anchor. The text is free -- "annoyed", "trying not to laugh", "dawning horror" are all fine. There is no list to choose from, because the shading is the point. Each clip takes its own shade of one colour chosen for the character, so a run of them reads as a run rather than as one long block.
Shaping the curve:
- Split a clip with the Split tool, exactly as you split a shot. Both halves say the same thing to begin with; you then change one of them. That is how a turn is authored -- cut "annoyed" in two and make the second half "angry".
- Double-click a clip (on a touch screen, tap it to select it and tap again) to edit it where it sits. The box asks three things left to right: whether this is a Mood or a Shift, what it says, and whether to keep it. Enter keeps what you typed and the tick does the same; Escape leaves it as it was, and so does the cross. Clicking anywhere else keeps what you typed rather than throwing it away.
- Drag the split point between two moods. Each one carries a diamond, exactly as a cut between two shots does -- and like shots, you never drag a mood clip itself or take hold of its ends. One mood gives exactly what the other takes, so the pair always covers the same span, and the drag stops when either of them would run out. The start of the scene and the end of it carry no diamond: the first mood starts where the scene does and the last runs to its end.
- Delete merges a clip into its neighbour rather than leaving a hole. The last remaining mood in a scene cannot be deleted; delete the track if you want the curve gone.
- A mood cannot be slid along its track on its own, because it runs up against the ones either side. Move the boundary, or split to add one.
- Inserting or removing time moves moods with everything else, and never opens a hole in the curve.
Shifts. A clip is either a Mood -- a feeling held over its span -- or a Shift, which is what an actor plays when a feeling turns. The pair of buttons at the left of the editor is which one this clip is; press Shift to make it one. A shift means "moving from the mood before this to the mood after it", and it stores neither end, so changing what either neighbour says keeps the shift correct. It is drawn as the clip split along its diagonal, with no words on it: it has none of its own. A clip at the very start or the very end of a scene cannot become a shift -- there is nothing on one side to move from or to -- and the button says so rather than leaving you with a half-formed one. While it is a shift the field shows the wording it currently resolves to, greyed, so you can see what it amounts to. Press Mood to turn it back, and the words you had typed come back with it.
Reading the curve. Open the track from the (i) in the track column. A mood track's panel is its own -- it has no name to edit, no volume and no ducking, so all it offers is the curve, a way to delete the track, and Cancel. The curve is written out scene by scene, timed from the start of each scene:
0s: neutral
3.2s: annoyed
3.7s: starting transition from annoyed to angry
4.2s: becoming angry
4.7s: angry
A mood gives one line, at the moment it takes over. A shift gives two: the moment the turn starts, and the moment it has mostly arrived.
A track added to an episode covers the scenes that exist at the time. A scene created afterwards has no mood in it until you add one.
Seeing what something is
The wastebasket button in the toolbar lights up as soon as something is selected, and Delete on the keyboard does the same thing.
With the Select tool, double-click a shot or a clip, or press the white (i) at the right end of its bar, to see what it is: which scene it belongs to, when it starts and ends, how long it is, and for a clip its track, level, fades and where its sync point sits. It also explains the two things that surprise people most -- that a shot's length is set by you rather than by its footage, and that a clip is anchored to its scene, which is why editing shots never drags it out of place.
Rendering happens at the shot and scene level -- see Shots for render controls. To render a whole episode, ask the Assistant to render this episode: it shows the shots and a cost line, asks Local or Cloud, then makes each shot's storyboard, rough clip and finished clip from its own visual style's recipe picks, scene by scene. See Visual styles.
Media
The episode's own Media Gallery -- the home for anything that belongs to the whole episode rather than to one scene or shot, such as a music bed or a score that runs across scenes. It behaves like every other gallery, and clicking the episode in the Browse Media Files dialog shows the episode's own media (expanding it to reach its scenes and shots is the same click).
Performances to record
The report of every shot you have said is driven by a performance you record, and that does not hold its recording yet -- your own acting, your face or your voice. It is listed by scene, and each shot shows what to act (its action directions or description), who is in it and how long it is. Add the recording uploads a video or audio file into that shot's media, marked as a performance recording; the shot then leaves the list, and its clip is made from the recording with your pick for shots with a performance recording (see Visual styles).
You say which shots are driven by a performance in Record performances, in The Shoot (and on the shot's Properties page, Performance card). Only scenes with characters are asked. Whether a shot is recorded is read from its media, so a recording added any other way takes the shot off the list too.
Progress
The Progress view is the episode's heat map: the ten production phases as columns, everything the episode is made of as rows, and a green / white / grey cell for each pairing. It is the fastest way to see which phase is nearly done and which has barely started, and every cell is a link to the work it stands for. See Progress for the full reference.
Guide
The Guide view is the guided, step-by-step way to work through the episode's production plan -- one step and one group of fields at a time, with AI assistance close at hand. It is the primary place to build an episode out. See Guide for the full reference.
Exporting the episode
When the episode is assembled, the Sequencer can turn it into a single video file. Open the Render menu at the top right and choose Export Episode Video (it exports at the episode's Finished size; change it in the toolbar first): every shot -- the chosen clip or storyboard still, in scene order, with the dialogue and music mixed in -- is rendered and combined into one video, which downloads to your computer to watch and share. If some shots have changed since they were last rendered, you are asked to re-render them first. Shots that have no chosen image or clip yet appear as blank frames, so you can still export a rough cut and fill the gaps later.
To take just the sound, choose Export Episode Audio right below it. Nothing is rendered: the episode's dialogue, music and effects are mixed -- with their levels, fades and ducking, exactly as for the video -- and downloaded as a WAV file, which is the same soundtrack the video carries and lines up with it from the start. The item is greyed out when the episode has no audio clips.
The exported file runs from the first cut to the last one, and its sound is exactly as long as its picture. Anything placed past the end of the final shot -- a line that overruns, a music tail -- is not in the export, because there is no picture out there to play it over. If something needs longer, give the last shot the time and it comes with it.