September 24, 2026
Scriptri 0.6 lets you see an entity as it stands at a chosen point in the story, tie continuity records to the exact manuscript passages behind them, and keep your own summaries for each scene and for the whole project.
Scriptri 0.6 is about getting from a recorded fact back to the place in the manuscript that supports it, and about seeing what is recorded at the exact point in the story where you need it.
Scriptri has always let you write down what characters know, believe, feel and want, and where they are. Those notes are attached to scenes. That leaves two questions writers may ask constantly while writing. What is true about this character right here? And which sentence in the manuscript makes me think so? Before 0.6, answering the first meant working through a filtered list, and answering the second meant remembering or re-reading the manuscript.
0.6 gives each question a direct answer. It also adds two places for the writer's own summaries, one for each scene and one for the whole project, so that you can find and recall your own story without rereading it.
Everything in this release is written by you. Snapshots are assembled from notes you recorded. Evidence is text you selected. Summaries are text you typed.
A note that says "Henry believes Lauren is safe" is only useful if you know when Henry believes it. Scriptri records story position by scene, and it always has. What it did not do was hand you one composed answer for one entity at one position. You could set up At This Point to show carried-forward notes, then scan the groups yourself.
The first two features in this release work on that gap from opposite ends. Snapshots compose what is recorded for an entity at a chosen point. Manuscript evidence lets a record point at the exact words behind it, and sometimes at an exact position inside a scene.
A Snapshot shows one entity as it stands at a chosen point in the story, built only from the notes and events you have already recorded.
Here is the kind of result it produces, from After the Choirs Fell Silent, the manuscript I am working on right now:
Henry Gearhart, before The Comedown At Kalsett Tower. Believes Lauren is probably safe, but remains emotionally unsettled by her rejection. Knows the attack involved at least ten combatants. Has not yet learned that disciplinary action is likely.
The Snapshot sits beside the Profile as its own tab in the Entity workspace. It is also what you see in the Floating Reference Viewer. If you click an entity mention in a scene and choose View, the panel opens on that entity's Snapshot, already anchored to the scene you are in and set to Before. That is one action from the manuscript to "what is true about this person at this point."
Every Snapshot names its position: a scene, and either Before or After it. Before shows the state as it stood going into the scene. After includes what the scene itself recorded. When you open a Snapshot from anywhere other than a manuscript mention, you choose the scene yourself. Scriptri does not guess one on your behalf, and if the scene it was pointed at no longer exists, you are asked to choose again.
Inside each topic, the note in effect at that point appears in full. The one before it and the one after it are a click away, and a marker shows when something changed in the anchor scene itself. Each item has a source action that takes you to the scene it came from.
Snapshots work for every entity type, including custom ones. A Place or a Faction gets the same treatment as a Character. When little has been recorded, the Snapshot says so plainly instead of showing an empty box.
A Snapshot holds no data of its own. It is a reading of your existing notes, so it cannot drift out of date, and nothing in it is inferred or generated.
Building the location section exposed an inconsistency, and the fix changes an existing view. In At This Point with the mode set to Before, a location event recorded in the anchor scene itself is no longer used for that entity's location. Scriptri falls back to the most recent earlier event, or to unknown. Notes already worked this way, so Before now means strictly before the scene for both.
Scene-level anchoring is often too coarse. Suppose Henry is at Kalsett Tower for most of a scene, leaves near the end, and is at his apartment when the next scene opens. A note attached to the scene cannot say when in the scene the change happened. And a note can quote a supporting line, but two notes that rely on the same paragraph had no way of knowing they shared it.
Manuscript evidence addresses both.
Select some text in a scene and choose Track passage from the selection menu. From there you can:
Your prose does not change. A small marker appears at the end of the tracked range, and that marker is the only thing you click. The same passage can support several records, and a record can cite several passages. Tracked passages can overlap or nest, and each stays independently trackable. If you select a range that is already tracked, Scriptri reuses that passage instead of creating a second one.
The word for this is evidence. It is deliberately not "fact." Prose can misdirect, remember wrongly or belong to a character who is mistaken. A tracked passage is a pointer to the words that led you to write a record, and the record stays your interpretation.
This is the distinction that makes the feature useful, so it is worth stating carefully.
Citing a passage means the record is supported by those words. Nothing about the record's timing changes.
Marking a passage as the record's effective position is a separate, explicit choice. It says that this change happens at this point in the manuscript. Scriptri asks you to confirm it in words that name the record. The position used for ordering is the start of the tracked range, since that is where the described change begins. Extra passages on the same record remain evidence only, and you can clear an effective position later without deleting the connection.
With an effective-position passage for a change partway through a scene, Henry's departure from the tower is placed where it happens in the text. A Snapshot or At This Point view will order several such changes correctly within the scene, instead of treating them all as happening at the scene boundary.
A passage's position never becomes the story time of everything it mentions. A Timeline Event keeps its own normalized time even when the same passage, used elsewhere as the position for a knowledge note, describes that event in a briefing scene.
Clicking a marker opens the Evidence Inspector. It shows the excerpt, the state of its source, and every record connected to it, with an action to add another connection.
The other direction works too. Every place that shows a State Note, Timeline Event or Location Event now includes a Supporting passages list. It shows each cited passage with its source status and a Go to passage action. That action scrolls the manuscript to where the passage currently sits, even if you have edited the scene since, and briefly highlights it. If Scriptri cannot resolve a passage to exactly one place, it says so where you clicked and does not jump somewhere approximate.
A project-wide Evidence panel lists every tracked passage. You can search it by quotation, scene or connected-record text and sort by connected record, source-scene order or most recently changed.
Excerpts also appear in Snapshots, At This Point and the entity thread timeline, so the words behind a note are visible where you are reading it.
A control in the editor toolbar sets how tracked passages display: annotation markers only, one of four highlight colors, or hidden. This is a display preference for your browser. It is never saved into the manuscript or the evidence records, and another browser starts with the default. Amber stays reserved for Lore links, and a Lore link always takes visual priority where the two overlap.
Manuscripts change constantly, and tracked passages are built to survive that without pretending to more certainty than they have.
Tracked passages travel with the project through export, import and copy, alongside the scenes they belong to.
Each scene now has an optional summary that you write. It is for recall and navigation: a line or several paragraphs saying what the scene does, or what you intend it to do. It can describe finished prose or a scene you have not written yet.
The Scene summary row sits in the Write header under the scene's point-of-view, time and location rows. It is collapsed by default, with a small dot to show a scene has one. Open it, type, and it saves as you go. Type @ to insert an entity chip. Chips are renamed along with the entity, so a summary does not go stale when you rename a character.
A summary is separate from the manuscript. It does not count toward word count and does not appear in a manuscript export. It is also separate from continuity records: it is not a note and not evidence. Scrivener sync never overwrites it, since it belongs to Scriptri.
Summaries are also exported, snapshotted and backed up with the project.
Projects had one short description, sized to identify a project on its card. That is a different job from saying what a whole manuscript is about, so 0.6 adds a Project synopsis for the longer account.
The fields are deliberately separate:
Both are edited in Project Details. Open it from the pencil on a project card, from the project name in an open project, from Data, then Project Details, or from a Quick Reference result. Quick Reference has a new Project group that matches the name, short description and synopsis, and opens Project Details at the field that matched.
The dialog has a view mode and an edit mode with explicit Save and Cancel. If you have unsaved changes, closing it any way at all (Escape, clicking outside, Close, the browser Back button or a page refresh) asks first, with Keep editing as the default. The synopsis and short description both accept @ entity chips, which follow renames. In an open project, activating a chip opens that entity in the Reference Viewer.
The synopsis stays out of readable manuscript export. It is part of the complete project export, snapshots and backups. Scrivener import and sync leave it alone.
Custom Entity Types in 0.5 let each type have a color and an icon. Entities picked up the type's color automatically, but an entity without its own icon showed nothing, even when its type had one.
In 0.6, an entity with no icon of its own shows its type's icon. If you give an entity an icon, that icon wins. This appears in entity lists, the entity header, entity selectors, the @ menu and the entity chips in the Timeline, Time and Context views. Entity mention chips in the manuscript itself are unchanged.
Changing a type's icon updates every entity that inherits it immediately. Changing an entity's type changes the icon it shows, and an entity's own icon stays put through a type change. When you edit an entity, a short line under the icon says whether it is inheriting or custom, and a Use type icon button returns a custom icon to inheriting. If a type has no icon, nothing changes for its entities.
Two pieces of work are underway for the next release.
The first is a rework of Investigations into story topics. A topic ties a set of related data together under an explicit name, and that grouping can be used as an investigation surface. The branding and the details are still being revised, so expect the name and the shape to change before it arrives.
The second is a consistency pass on how entities are tagged, along with refinements to popovers and a move of some dialogs into floating panels. Quick Reference is the first example.
Nothing else is firmly locked in yet, though I expect that release to hold more than these. A much larger update is also in development, further out than the next release. None of this is a schedule, and priorities will keep following real manuscripts and the feedback of writers who use Scriptri. If something in 0.6 gets in your way, or Snapshots or evidence do not behave the way you would expect, the feedback form on the home page is the fastest way to tell me.