Context
Start with the Story you want to write. Its Working Context selects Story, Storyform, and Persona for the current task and gives you access to Ideas. An Individual Project remembers that selection, so you can work alone without first arranging a larger collaboration. Read What You Work On for the introduction.
Story Context holds that Story's saved material, including its Story Guide and Story Expert work. This page explains how Narrova uses the selected sources and conversation memory. Open Project button → Working Context to inspect the available tabs, and use Collaboration and Projects when you need more organization, saved presets, or shared work.
Role is Who you work as: Individual, Team, or Group. Project is What you work on. Working Context is How you approach it through your selected Story, Storyform, and Persona, with access to personal Ideas. Workspace is Where you work: Plot Progression, Story Expert, Narrova, Storyform Builder, or Subtxt.
Working on identifies the Story or exact Storyform you intend to develop. Working Context identifies the references and guidance you want to use. A solo writer will usually select the same work in both roles; collaborators, writing agents, and episode or season development can need different reference material. Reading a series bible should not make it the destination for an episode's accepted changes. See what you work on and how you work, including the current differences between platform controls.
Narrova lets you build with intent by combining durable Context with files attached to the current conversation. Each layer has a different scope, so responses stay sharp, relevant, and true to your current objective.
Context types at a glance
- Story Context keeps broad material that applies across many ideas: world lore, character bios, visual references, tone guides, research notes. Think of it as your shared universe bible.
- Storyform Context holds the structured pieces of a single Dramatica Storyform: Throughline workpads, signpost explorations, encoded beats, thematic breakdowns. It is the “structured spine” for one story.
- Ideas in an Individual context, or My Contribution → Ideas in a shared context, is your personal stream for thoughts and quick memos, without a required category, Story, or Storyform. Ideas is a separate collection selected by the Working Context; it can belong to that context, a Team, or a Group. See Ideas.
- Files holds files for the current discussion, separately from Ideas and lasting Story or Storyform libraries.
- Persona supplies the perspective or instructions for assistance. It can use defaults, a selected Persona, or No persona; it does not change access permissions.
Solo and team workflows
If you are working one Storyform inside one story, Storyform Context will usually be the most useful place to stay. Put the cast conflicts, locations, scene notes, encoded Storypoints, draft material, and Storybeat work there so Narrova can keep answering from the structure you are actively developing.
Use Story Context when the material is bigger than one Storyform: worldbuilding, shared continuity, recurring locations, series rules, tone guides, research, character histories, and anything you keep repeating across conversations or Storyforms. It is also helpful when you are not sure about the Storyform yet and want Narrova to keep the larger creative frame in view while you explore.
For development teams, a shared Story Context can hold the story-world material. A series Team might keep the franchise bible, world rules, character canon, and season-level continuity in a Team-owned Story, while episode Groups work in their own Stories and Storyforms. Link those resources in season or episode Projects, then choose the specific Story and compatible Storyform needed for the task. Project links do not merge source material or grant access to it. Each episode can technically contain more than one Storyform, but most episodes are cleaner with one complete Storyform unless the runtime or format gives the story enough room to carry multiple complete structures.
The working mental model
Shared Working Contexts show Shared Persona and Shared Ideas panel headings. Their On/Off switches retain the selected Persona or collection, so you can include it again without choosing it again. The Persona switch leaves inherited Team and Group guidance active; the Ideas switch controls retrieval, not collection access or editing. Individual contexts continue to show personal Persona and Ideas.
Working Context selects what you are using now; it does not replace the places where the work is saved. Keep these parts distinct:
-
Project / Working on identifies the intended work. Where separate Save to controls are available, the exact Story and Storyform destinations determine where requested or accepted changes belong. Selecting them is not permission to save every brainstorm.
-
Working Context selects Story and Storyform. In an Individual context, its Persona and Ideas tabs use your personal settings. In a shared context, those tabs select context resources, while My Contribution separately remembers your personal Persona and whether to include your private Ideas.
-
Conversation records the discussion and immediate instructions for that session.
-
Memory preserves meaningful decisions and continuity facts in the appropriate conversation, Story, or Storyform scope.
Let the Story and Storyform hold lasting material, keep each discussion focused, and inspect Memory when you need to confirm what Narrova saved. A saved Working Context belongs to an Individual, Team, or Group; a Project can be related to it, but is not required. A Project can remember a selection without turning its linked work into a single source library.
What belongs where
-
Use Project Files to collect references before choosing a Story or Storyform. Open the Project's Overview → Files. With no Project selected, new files default to My Files. Creating work later keeps these originals in place.
-
In Files, All combines Project/My Files and the selected work's Story/Form files with source labels. Custom Working Context references do not change upload destinations. Existing conversation files stay accessible; they are not moved automatically.
-
Use Story Context for broad canon: character bios, world rules, tone references, research, visual inspiration, and notes that should stay true across multiple versions of the story.
-
Use Storyform Context for structure-specific canon: Throughlines, Signposts, encoded beats, thematic notes, Player mapping, and transcripts that belong to one specific Dramatica Storyform.
-
Use the live Conversation for directives, decisions, and the current task. Think of it as a work session, not your project's permanent memory.
Worldbuilding and codex files
Many writers use a “codex” file as a personal world bible: history, factions, magic systems, character biographies, continuity rules, visual references, voice notes, and prior-story events. Narrova can use those files well, but the file’s location matters.
Put worldbuilding files in Story Context when they describe the larger creative universe. This includes setting rules, recurring locations, character histories, metaphysics, research, cultural references, and continuity that should apply across more than one Storyform.
Put current-structure files in Storyform Context when they belong to one exact Dramatica structure. This includes the selected Main Character, Influence Character, Protagonist, Antagonist, Throughline work, Storypoints, Signpost plans, Storybeat notes, and draft pages for that Storyform.
If you are writing a series, organize prior events as history, not as instructions to repeat them. A file named Prior Events - Season One.md tells Narrova what happened before. A file named Current Storyform - Episode Two.md tells Narrova what structure it should help develop now.
Writing use: Keep broad canon in Story Context, keep current structural decisions in Storyform Context, and use the current conversation to say what you want Narrova to do next.
Role-neutral character notes
Character documents travel better when they describe the person before assigning a Dramatica role. A durable character file can say what the character wants, fears, hides, knows, believes, and does when plans fail. The Storyform can then decide whether that character is functioning as Main Character, Influence Character, Protagonist, Antagonist, or another Player in the current structure.
Use role-specific language only when the file belongs to that Storyform. For example:
- Story Context:
Walburga - Character Dossier.md - Storyform Context:
Walburga as Main Character - Current Storyform.md - Story Context:
Series Timeline - Prior Events.md - Storyform Context:
Episode Two Signposts and Storybeats.md
A better way to work
Start a new conversation whenever the objective changes, even if you are still working inside the same Storyform.
For example, during Encoding you might use:
- one conversation for Objective Story Players
- one for the Main Character's personal conflict
- one for the Influence Character's challenge
- one for Relationship Story tension
- one for a specific Storybeat or Signpost cluster
This does not break continuity as long as those conversations stay attached to the same Storyform Context. The Storyform carries the structural spine forward, so the new conversation does not need to inherit the entire old thread to stay aligned.
Memory
Narrova Memory is the durable ledger behind Context. It records meaningful story decisions, continuity facts, structural choices, and recovery notes so Narrova can carry them forward without forcing every future conversation to reread the entire chat history.
In plain language: Context tells Narrova what source material is available. Memory tells Narrova what has already been decided.
Narrova uses three memory scopes:
- Conversation Memory belongs to the current thread. It keeps track of local decisions, recent instructions, corrections, and work-session outcomes.
- Story Memory belongs to the Story. It carries broad continuity, world rules, recurring characters, series history, and decisions that should apply across multiple Storyforms.
- Storyform Memory belongs to one Storyform. It carries structural and development decisions for that exact interpretation of the story: Throughlines, Storypoints, Player mapping, Signposts, Storybeats, and accepted revisions.
These scopes work together. The current conversation gives Narrova short-term working context, Story Memory carries the larger creative world across conversations, and Storyform Memory keeps one structural version coherent while you develop it over time.
Memory citations
When Narrova commits something important to memory, it can show a memory citations control under the response. Open it to see which memory ledger changed, the line range that was updated, and a short summary of what was committed.
The citation labels tell you where the memory lives:
CONVERSATION-MEMORY.mdfor the current Narrova threadSTORY-MEMORY.mdfor Story-level continuitySTORYFORM-MEMORY.mdfor the Storyform identified by that citation
Click Open memory lines on a citation to open the side panel for the exact ledger excerpt. Use this when you want to confirm what Narrova saved, check whether a decision went into the right scope, or inspect a continuity correction.
Continuity modes
Some Storyforms continue directly from earlier events. Others use the same world or familiar characters with a cleaner slate. Narrova Memory supports that distinction so prior story material does not get repeated when you do not want it.
When starting or attaching a Storyform, choose the continuity mode that matches the job:
- Continues the main story uses prior events and memories as active continuity.
- Same world, clean slate keeps reusable world context while avoiding old plot events, player history, and relationship developments as required beats.
- Shared characters, reset events keeps character familiarity while treating prior events and relationship outcomes as background rather than current-story instructions.
Writing use: before opening a spin-off, decide whether the new Storyform is a sequel, a clean-slate story in the same world, or a reset with familiar people. Set the continuity mode first, then ask Narrova to develop from that premise.
Changing what Narrova remembers
You can edit memory through conversation. If a memory is wrong, outdated, too broad, or stored in the wrong scope, tell Narrova what to change.
Use direct instructions:
- "That memory is wrong. Maya and Jonah are siblings, not former lovers."
- "Forget that Evelyn betrayed the crew. We changed that."
- "Move that memory to this Storyform only, not the whole Story."
- "Treat this spin-off as a clean slate. Do not use the old events."
- "Update the Story Context: the city is recovering after the blackout, not still in collapse."
Narrova should identify the affected scope, revise or supersede the ledger entry, and use the corrected version going forward. When the correction is important, Narrova can cite the updated memory so you can verify the change.
Common mistakes
- Treating Memory as a hidden black box. Correction: open memory citations when they appear and inspect the lines Narrova saved.
- Leaving broad world facts in Conversation Memory. Correction: ask Narrova to move durable lore into Story Memory.
- Letting one Storyform inherit events from another by accident. Correction: set the Storyform continuity mode to Same world, clean slate or Shared characters, reset events before developing the spin-off.
- Assuming an old memory is permanent. Correction: tell Narrova exactly what changed, what to forget, or which scope should own the revised memory.
Check your current Context
Begin with the Story you want to develop and the exact Storyform version, if the task needs one. Check Working on and the save destination separately from Working Context, which selects the Story and optional Storyform references, Persona, and available Ideas. An Individual Project remembers that selection; you can keep writing on your own without first organizing a larger collaboration. See What You Work On for the introduction.
The Project's remembered selection and an existing conversation's saved attachments are related but separate. Before continuing an older conversation, inspect its own Context as well. Selecting another Project does not rewrite every conversation in your account.
The Project button and Working Context
On the web, open the Project button. The Project type and title appear first, followed by Working on, then the separate Working Context name and source grid between light rules. Clear Project and Manage Project share a row; Switch Project is at the bottom. Open the Context grid or a saved Context card to inspect the exact reference titles.
Manage Project lists the selected titles on separate lines under Working on above Overview, Access & Roles, Development, and Goals. This summary is not a Context chooser. Use Overview → Working Context or the pop-up's Context row to change references. In Narrova versions with Save to, inspect those exact destinations in Project as a separate choice; the web preview alone does not establish independent save routing.
Use the Story or Storyform tab to select the reference material you want the task to use. Opening a resource's manager lets you inspect Overviews, Development, and Context Materials. Inspection alone does not select a different resource; finish the grid's selection flow to apply it.
You can also open Project button → Manage Project → Overview → Stories & Storyforms, open a linked resource, and choose Use in Working Context. Choose Use Working Context to apply it, or Use and Save… to continue into saving a preset. This keeps your Persona setting. Changing the Story or Storyform starts without the previous conversation's files; saved presets change only when you explicitly save.
Manage Project opens the organizational controls for Overview, Access & Roles, Development, and Goals. Clear Project removes the active Project selection while keeping compatible working material. Clear in the grid clears the Story and Storyform selections while preserving saved work and the Persona preference. Neither action deletes the Story, its material, or conversation history.
Context inside a Narrova conversation
Narrova's composer has its own Context control for the current discussion. Its active sources can include Story, Storyform, and conversation files; on a narrow screen, the summary can collapse to a source count. Open it to inspect the actual attachments rather than assuming an older conversation follows the Project's latest selection.
On the web, + → Set Context also opens the conversation dialog. + → Add files and images opens the composer attachment flow, while + → New Story starts Story creation. In the Context dialog, Files contains the conversation's immediate files. The Story and Storyform tabs lead to material saved with those resources.
These instructions describe the current web controls. Native app layouts and save labels can differ. Follow the native app's explicit apply or confirmation action, then check the selected titles before continuing; closing a sheet alone does not establish that a pending change was saved.
Manage Context
The Context grid offers Story, Storyform, Ideas, and other tabs depending on where you open it. Story and Storyform identify the selected reference material; Persona guides the approach to assistance. The work receiving accepted changes may be the same records or, where separate destinations are supported, different ones. Ideas is a personal place to capture thoughts without assigning them to a Story or Storyform. Files holds conversation-specific files and may show read-only references from the selected Working Context.
Opening a tab changes the panel you are viewing. Review the selected Story and Storyform and the relevant Persona before applying changes. Capturing an Idea does not require selecting a Story or Storyform.
| Tab | Use it for | Example |
|---|---|---|
| Story | Saved material for the selected Story. | Character backgrounds and research. |
| Storyform | One specific narrative structure and its development. | Throughlines, Storypoints, and Storybeats. |
| Ideas | Personal thoughts and quick memos. | A question or fragment you want to remember. |
| Persona | The perspective and instructions used by assistance. | Your preferred editorial approach. |
| Files | Files for the current discussion. | Today's scene excerpt. |
Story
Select Story to choose the broader narrative material Narrova should use. Choose a Story from the available list, or use Open Another to find a different one. A selected Story can narrow the available Storyforms to compatible structures.
The selected Story's material manager offers Overviews, Development, and Context Materials. Keep character biographies, setting rules, research, and other lasting material with the Story so it can support later work. Save structural assignments with the Storyform that uses them.
Turning Story off removes it from the pending selection; it does not delete the Story or its files. Review the remaining Storyform and its compatibility before applying the change. To browse beyond a filtered pairing, clear the incompatible source and select again.
Storyform
Select Storyform to choose a particular structural interpretation. Use Open Another to search available Storyforms. Where permitted, Create Empty Storyform creates one attached to the selected Story; the new structure still needs your development and choices.
Story and Storyform can be included independently, but they must be compatible when both are selected. Check the intended version before structural work, especially when one Story has alternatives. Turning Storyform off keeps a selected Story; turning both off is supported, although a Workspace can require one of them before its tools open.
A Canonical example · Read-only or Reference Storyform · Read-only label means you are using reference material. You can inspect the example, but its document controls do not let you edit or upload to that reference.
Save or discard Story and Storyform changes
Here, saving Working Context means saving its source configuration. It does not save proposed Storytelling, change Project save destinations, or move a conversation. Those actions have their own controls and confirmation.
When editing a Working Context on the web:
- Open Stories & Storyforms → Working Contexts and select a saved Context.
- Review its Story, Storyform, Persona, and Ideas in the grid. Conversation files are managed separately.
- Choose Update Working Context to save changes to an editable Context. Use Save as Working Context in the manager to create a separately named copy.
- Choose an Individual, Team, or Group audience for a new Context. A Project is optional; when offered, Relate to [Project] associates it with the selected Project.
- Reopen the Context to confirm the exact titles and sources.
Your Project remembers your own current selection. Editing it does not automatically update a shared preset or switch another writer's active Context. If you switch saved Contexts with unsaved changes, review Keep Editing, Switch Without Saving, or Save Changes deliberately. Read saved Working Contexts for the complete save workflow and sharing limits.
When changing an existing conversation's Context, choose Done to apply the staged Story and Storyform choices. Cancel discards those pending choices. Conversation Persona changes and file operations save through their own controls; cancelling the dialog does not undo operations already completed there. A Persona selected in Working Context is instead staged with that Working Context, rather than changing the account default immediately.
Workspace changes save supported pending writing before loading the selected Context. Resolve a save error before switching and confirm the destination afterward. Files, Notes, and other material already saved to a Story or Storyform remain there when you change or clear a selection.
Files
Open Files in the Context dialog for the current conversation's files. These are separate from the lasting libraries under Story → Context Materials and Storyform → Context Materials, and from your personal Ideas.
Choose Add Files or drop a file into Files. Where supported, the app prepares a conversation for those files; otherwise, start or open a conversation first. Wait for processing and indexing before expecting an uploaded document to inform Narrova's answer. Use the available visibility controls to include or hide individual files.
Use the file's three-dot menu for Add to Story Context, Add to Storyform Context, or Delete, when available. Adding a file to Story or Storyform Context creates a copy and leaves the conversation file in place. You need Write or admin access to both the source conversation and the destination Story or Storyform. Apply the intended conversation attachment first, then check the resource named by the file-copy action. A pending attachment choice has not changed that action's destination. These file-copy controls are separate from Narrova's Project save defaults. Verify the copy in the named resource's Context Materials.
File changes take effect separately from the grid's staged narrative selections. Closing or cancelling the dialog does not undo an upload, visibility change, copy, or deletion already completed. Enabling retrieval or uploading may require AI consent. Hiding an included file does not require enabling AI sharing.
Personas
A Persona supplies saved instructions for Narrova's approach and explanations. Its scope depends on where you select it. Create and edit the underlying Persona in Preferences, then choose how the current work should use it.
In Project button → Working Context → Persona, the panel identifies Working Context Persona. Select an available named Persona or No persona; a Context can also retain an inherited default selection. The choice is saved with this Working Context and does not activate a different account default. Keep personal Persona choices in an Only me Context when colleagues cannot access them.
In an existing conversation, Context → Persona identifies Persona for this conversation and offers:
- Use defaults: inherit the Story default when one exists, otherwise the account default.
- A named Persona marked Conversation override: use it for this conversation.
- No persona → Override all defaults: use no Persona in this conversation, even when defaults exist.
Conversation Persona changes affect the next message and preserve its history. A Story default chooser appears when you have permission to set it; Use account default removes that Story-specific default. If an existing conversation offers Apply updated [Persona name], use it deliberately to adopt the saved update.
An inactive account Persona does not prove that an existing conversation uses no Persona: its Story default or conversation override can still apply. Likewise, selecting a Persona for Working Context does not change Team membership, sharing, or edit permissions. Team and Group Instructions have their own scope in shared conversations.
What carries across workspaces and conversations
A Project remembers each writer's own Working Context. Switching Projects restores your selection for the destination Project; changing Workspaces uses the selected material for the next task. Inspecting a Story or Storyform does not change the current Project or choose it as Context. Use the explicit Open in…, Set as Current Project, Use in Working Context, or save-destination action offered by that surface, and check the result. Opening references does not automatically refile an existing conversation.
An existing conversation retains its own saved attachments and history. Selecting global Working Context does not rewrite that older conversation, copy its files into a new discussion, or merge Stories. New work can use the same lasting Story and Storyform sources without importing every conversation file from the previous session.
Workspace requirements still apply. Story Guide needs a Story and keeps its Story-specific workbook and discussion. A Storyform-based Workspace needs a Storyform. Clearing a required selection does not create a missing Story, structure, or workbook; choose the intended material before continuing. See Choosing a Workspace for handoffs.
Ask Narrova to change global Context
Give an explicit instruction with an unambiguous target, such as “Switch to Project Harbor,” “Set Working Context to Story Juniper Road and Storyform Draft Two,” or “Clear Working Context.” Mentioning a Story during creative discussion is different from asking to switch to it. An ambiguous title needs a choice before the app can act.
For a typed request, review the proposed Project and Working Context in the app, then choose Confirm Context when you intend the change. Wait for confirmation and inspect the result before doing dependent work. A request that is cancelled, expired, or unconfirmed does not establish a successful switch. These controls select working material; they do not automatically change Storytelling, structural choices, ownership, or access.
Use the app's Project controls in shared conversations where private Project-selection tools are unavailable. If your chosen Story is not yet organized in a Project, an offered Individual Project creation is an organizational step for that Story, not a requirement to begin a broader collaboration.
Changing Context during Voice Mode
The current web Voice call can continue while its Working Context changes. The app preserves the call's transcript and uses compatible working material for subsequent requests. A completed switch does not rewrite an older conversation's attachments or convert the conversation files into a Story Guide workbook.
Give an explicit spoken instruction and wait for the app to acknowledge the resulting selection. Spoken authorization can allow the voice path to apply the change without the typed request's confirmation dialog. A typed Context request still uses its visible confirmation, including while Voice is minimized. Check the selected titles afterward; an uncertain or failed acknowledgement is not proof of a completed switch.
Native interfaces and recovery actions vary by app version. Follow the action the app presents and verify its saved selection before continuing. See Talk to Dramatica for access and session guidance.
Context usage is a separate control
The optional context-usage meter measures the current conversation's capacity. It does not tell you which Story, Storyform, Persona, or conversation files are selected. Enable Show context usage in Preferences → Narrova Context to display it.
Use Narrova Settings for intelligence and response verbosity, Preferences for those settings in detail, and Narrative Efficiency for conversation headroom and compaction. A capacity meter and a source selection answer different questions.
Troubleshooting Context selection
When a choice looks wrong, inspect the Project, intended save destinations, and actual selected reference sources separately. Then check the current conversation's Context if you are resuming an older discussion. Keep saved work in place while correcting the selection.
| What you see | What to check or do |
|---|---|
| A different Story opens after a change. | Reopen Working Context, finish Continue and any confirmation or save flow, then verify the exact Story and Storyform titles. |
| A Story or Storyform is missing from the list. | Clear the search and remove an incompatible other source. Check access and retry a failed catalog load. |
| A save error blocks a switch. | Resolve the current writing or Storyform save error, then retry. Do not assume the destination loaded. |
| An older conversation shows another Storyform. | Inspect that conversation's own Context. Project selection does not rewrite every saved thread. |
| A saved Working Context is Unavailable. | Check access to every source, including Persona and conversation files. A shared preset does not share its private material. |
| Conversation files are empty or unavailable. | Check that the intended conversation has files. Story and Storyform uploads belong to their own Context Materials. |
| An upload is listed but does not inform answers. | Wait for processing and indexing, refresh status where offered, and check inclusion. |
| Some file changes could not be confirmed. | Refresh the files and inspect each state before retrying; a partial operation can leave a mix of included and hidden files. |
| Narrova still uses a different Persona. | Check Working Context, Story defaults, and the existing conversation's override. An account-default change does not replace every override. |
| A global request expired or changed during confirmation. | Inspect the saved selection and ask again. The app does not apply an old request as a fresh instruction. |
| Voice reports an uncertain Context change. | Check the app's acknowledgement and selected material. Resolve the offered recovery action before relying on the change. |
Conversation Files
Conversation Files are documents attached to one Narrova thread for focused reference. The conversation Context control presents them in Files; they are not a saved Working Context source. Use this scope for one-off excerpts, comparisons, and other material that should inform the current discussion without automatically joining a Story's lasting source library.
Open the conversation's Context, choose Files, and check that the intended files are included and ready. To keep a file beyond that discussion, use its menu's Add to Story Context or Add to Storyform Context action when available, then confirm the copy in that resource. See Organizing Context for keeping the libraries coherent.
Uploading documents
Uploading turns a document into usable Context. Narrova stores the file in the chosen Context, splits it into retrieval-friendly passages, and makes it available through the Context manager so you can control when it is visible.
For Story Context and Storyform Context, limits are based more on the overall document footprint than on the number of files.
- Dramatica: 2 GB document storage
- Dramatica Plus: 10 GB
- Dramatica Pro: 20 GB
That is why file count by itself is not a very useful measure. A few large drafts may matter more than many small notes, and many small files can work perfectly well if the overall Context stays coherent and intentional.
What uploading means
- Story Context upload adds a document that can support every Storyform in that Story.
- Storyform Context upload adds a document that supports only that specific Storyform.
- Conversation File upload attaches a file directly to the current chat for short-lived reference.
When uploading worldbuilding or codex files, use the optional Document intent field when it is available:
- World bible for broad setting, metaphysics, rules, factions, history, and series canon.
- Character profile for role-neutral character dossiers.
- Timeline snapshot for current chronology or continuity state.
- Prior story/events for events that already happened and should be treated as background unless you ask Narrova to continue them.
- Style guide for voice, format, expression, prose preferences, or translation notes.
- Other for reference material that does not fit the above.
Upload lifecycle
- Uploading – the file is transferring. You can keep chatting, but the file is not yet searchable.
- Uploaded – the file is stored. Narrova starts preparing it for retrieval.
- Indexing – text is extracted and split into passages. Once indexing finishes, the document is fully available for responses.
Managing document visibility
For conversation files on the web, open Context → Files. For lasting files, open the Story's or Storyform's Context Materials. Use the visibility controls offered by that view when the file is ready:
- On includes the file as retrievable reference after processing.
- Off keeps it stored in its current scope while excluding it from retrieval until re-enabled.
Use a file's menu to copy it into Story or Storyform Context or delete the conversation file when appropriate. Deleting that file does not delete a separate saved copy.
A Story or Storyform document's visibility belongs to that lasting library, so consider other conversations using it before changing it. Storyforms opened in read-only reference mode remain read-only; selecting one of your writable Storyforms as a Working Context source does not remove its edit permissions. Native apps can arrange these controls differently; verify the selected resource and file state in the view you are using.
Compaction in a conversation
Compaction keeps long conversations writable by replacing heavy prior context with a condensed version Narrova can continue from.
It does not erase your visible conversation history. It compresses active model context so Narrova can keep delivering focused, high-signal responses.
See Narrative Efficiency for the meter and controls that signal when compaction is useful.
Manual compaction
Run manual compaction with /compact in chat.
- Best when you want to keep the current thread but reclaim headroom right now.
- Narrova compacts from the active response-chain anchor in the current conversation.
- After compaction completes, Narrative Efficiency updates immediately with an estimated post-compaction value.
Automatic compaction
Automatic compaction runs when the thread reaches the hard context threshold.
- Before that threshold, Narrova can show a soft warning that compaction is recommended.
- At the hard limit, Narrova starts compaction automatically so the conversation can continue without a hard creative stop.
- The compaction notice in-thread marks running and completed states.
What compaction changes (before vs after)
Before compaction
After compaction
In both columns, your current prompt stays the same. What changes is the memory stack above it:
- Before: raw, full conversation history.
- After: a compact narrative snapshot that preserves what matters most for continuation.
Working inside a Story
Opening a Story reveals everything linked to that project.
Documents
Upload files you want Narrova to consider for every Storyform in that Story—casting sheets, location research, tone guides. Rename or remove documents anytime.
Conversations
Every conversation that references this Story appears here. Click a title to reopen it, favorite frequent threads, or use the overflow menu for more actions (see below).
Storyforms
Attach new Storyforms, set one as primary, or detach and reuse it elsewhere. You can still find every Storyform in the global index even if it is not currently attached to a Story.
Working inside a Storyform
A Storyform page focuses on the structure itself. Store encoded beats, progression notes, documents, and conversations that belong exclusively to this interpretation of the story. When you validate or adjust the Storyform, those updates guide every conversation that references it.
Conversation actions menu
The conversation overflow menu (⋯) exposes tools that help you keep context tidy.
Branch conversation
Use Branch conversation when you want to test a different creative path without disturbing the original thread. Branching duplicates the conversation up to the selected completed response, then opens a new conversation from that point. The original stays intact, and the branch gives you a clean place to make a different choice.
This is especially useful during Story development when Narrova asks you to choose between meaningful structural alternatives. For example, if Narrova asks whether the story should end in Success or Failure, branch at that response and try both. One conversation can explore the Success version, while the other explores the Failure version, letting you compare how each Outcome lines up with the Storyform, Throughlines, Storypoints, and Storytelling pressure.
To branch:
- Find the completed Narrova response where the alternate path begins.
- Click Branch conversation on that response.
- Confirm with Create branch.
- Continue from the new conversation that opens.
The branch carries forward the same Story Context and Storyform Context, so Narrova still understands the Story and Storyform you were developing. Conversation files attached to the original thread also remain available to the branch. To retain source material from either branch, save it in the appropriate Story or Storyform Notes, or use Upload into Context. Save approved component changes to the selected Storyform.
Each conversation keeps links to its related branch. The original conversation shows where it branched to, and the branched conversation shows where it came from. Use those links to move back and forth when comparing versions.
Writing use: when a structural decision could take the story in two honest directions, branch the conversation and let each version prove itself through Storyform alignment, character pressure, and scene consequences.
Move to Context
Use this when the conversation belongs somewhere else. Maybe you brainstormed under the wrong Story, created a new Storyform you’d rather use, or want to explore the same idea in a clean space. The Move dialog lets you review current attachments and select a new Story or Storyform without dismissing the modal. Once moved, the conversation immediately benefits from documents and structure that live in the new destination.
Upload into Context
Upload into Context saves the entire transcript as a Markdown document in the Story or Storyform named by that action. Check the exact destination before confirming; a transcript copy is separate from the destinations for accepted development. Narrova does not archive chats automatically—you choose what becomes long-term memory. When you trigger the upload, the conversation is stored as a document whose name matches the conversation title, so you can reference or reopen it later.
Keeping context manageable
Large chats are tempting, but smaller purposeful conversations stay cleaner, faster, and easier to steer. Let the Storyform carry continuity across sessions, and store durable knowledge as documents instead of leaving critical decisions buried in chat scrollback.
- Collect and label documents for each Story and Storyform. Short focused files (2–10k tokens) are easier for the model to retrieve accurately.
- Start a new conversation for each objective: locking Domains, testing a Relationship Throughline, exploring alternate Outcomes.
- Upload when finished. If a chat uncovers insights worth keeping, archive it with Upload into Context so future requests can build on it.
- Move conversations when their filing belongs with different work. Change reference Context when only the sources should differ, and check save destinations separately where offered.
Be deliberate about where information lives, and Narrova stays responsive while each response reflects the latest living version of your story world.