Where stories
meet the tide.
Your private writing studio.
PRIVATE WORKSPACE · AUTHOR ACCESS ONLY
PRIVATE WORKSPACE
Welcome back.
Unlock your writing space to continue.
Your session is protected by a server-verified access gate.
YOUR WORKSPACE
RES StudioTHE NEXT CHAPTER STARTS HERE
What are we writing?
Start fresh, bring an existing idea, or pick up where you left off.
DECLARATIVE CONFIGURATION
Niche pack library
Loading available packs…
Pack releases provide versioned writing guidance. Preview validates the pack structure; it does not call an AI API.
SERVER-SIDE MODEL ROUTER
Model settings
Checking provider configuration…
Catalogue checks only. No text generation is requested.
API keys stay server-side. Add keys as encrypted Cloudflare secrets: KFS_XAI_API_KEY, KFS_OPENAI_API_KEY, and/or KFS_ANTHROPIC_API_KEY. Then set each provider’s exact comma-separated allowed model IDs and default model in server-side environment variables. Never paste API keys into this screen, GitHub, or app-config files. Generation uses an explicit allowlist, input/output caps, daily request limits, and a monthly token budget. Token counts are not a dollar-cost ledger. Check billed spend in each provider console.
YOUR LIBRARY
Projects
Loading projects…
STORY BLUEPRINT
Story Format
Choose a project.
Choose the story's structural target. This saves planning parameters only; it does not generate prose or guarantee an exact final word count.
The format is versioned per project. Future outline generation can use these targets, but the current build does not automatically create chapter plans.
RES UNIVERSAL STORY DESK
Story Desk
Choose a project.
Build a structured concept from any creative dimension. Change formats by creating a linked variant, so the source premise and its history stay intact. This build saves concepts only; it does not generate prose or call an AI provider.
Concept lineage
Save or load a concept to inspect its lineage.
Saved revision history
Load a concept to inspect its saved revisions.
Format targets are planning guides, not guaranteed final word counts. A 6-in-1 concept records a six-story volume target; the individual story builder arrives later. Heroine Catalog mode records the recurring-heroine architecture; it does not impose a reading order. Drafting and automated QA are not triggered from this screen.
STRUCTURE BEFORE PROSE
Blueprint Workshop
Create a format-aware structure from the loaded concept, review the outline beats, then explicitly approve it. This is a planning scaffold, not manuscript prose. Saving and approval do not call AI.
Save a concept and build its blueprint to begin.
Blueprint versions
No blueprint versions loaded.
Approval locks this version as the approved structure. Saving changes creates a new draft version and leaves prior versions intact. A new format variant needs its own blueprint. Prepare in Writer Chat only fills the chat composer with an outline request. It does not send anything. API usage and possible charges begin only after you review the prompt and press Send message.
CRAFT, MEMORY & STORY LOGIC
Story Intelligence & Prose Craft
Choose a project.
Set the book’s craft rules, keep character voices distinct, and run focused story diagnostics. AI findings are saved as drafts; they never silently rewrite the manuscript or become canon.
Chapter / scene intelligence
Uses the latest saved text for the selected unit. Large chapters may need to be split. AI usage applies.
Saved story intelligence
BOOK ASSEMBLY & PUBLISHING OUTPUT
Book Assembly & Export
Choose a project.
Assemble front matter, chapters, and back matter without changing your saved manuscript. Export is a separate rendering layer.
Pre-export quality check
Load the manuscript to inspect the book structure.
Release readiness report
Run the checks, save a versioned snapshot, and review the release checklist before treating an export as ready. A saved report records the current check results, not a guarantee of marketplace acceptance.
Export book
DOCX is editable in Word-compatible editors. EPUB includes chapter navigation for Kindle Previewer / conversion checks. These exports are starting points, not a guarantee of KDP acceptance.
For KPF, import the DOCX into Kindle Create and export from there. Inspect the result in Kindle Previewer and KDP Previewer. Print page count, final gutter, headers, and page numbering should be checked after final layout.
REVIEW BEFORE CANON
Proofreader
Choose a project.
Run a focused review of a saved manuscript revision. The AI returns findings and a proposed full-text revision. Nothing changes in the manuscript unless you explicitly save the proposal as a new revision.
Review findings
The report is saved separately from the manuscript. Applying a proposal creates a new revision and marks it as a draft. You can inspect the result and approve it separately in Manuscript Workshop.
CANONICAL MANUSCRIPT
Manuscript Workshop
Choose a project.
Revision history
Select a chapter to see its revisions.
Draft Review & Continuity QA
Run deterministic checks on the text currently in the editor. Save a report snapshot if you want a historical record.
Uses the current editor text, including unsaved edits. Checks are advisory, not an AI proofread or a semantic contradiction detector. Findings never edit, save, or approve the manuscript.
No review has been run.
Saved QA snapshots
Snapshots preserve the report and findings, not the manuscript text. A saved report may become stale after edits.
No saved QA snapshots loaded.
Story Continuity Ledger
Record canon facts, then check whether their search terms appear in saved manuscript text.
No canon facts recorded yet.
Checks are literal text searches, not semantic contradiction detection. “No match” means the chosen terms were not found in saved manuscript text; it does not prove the story is inconsistent. The ledger never rewrites prose, approves chapters, or calls an AI model.
Saving creates a new revision. Approval is explicit. Restoring an earlier revision creates another new revision, so no previous text is silently overwritten. No AI request is made when editing, saving, approving, or restoring.
STORY STRUCTURE
Outline Workshop
Choose a project.
Draft chapter and scene beats, save versioned outlines, and approve the structure you want the writer to follow. Saving and approving do not call AI.
Continuity Context for This Chapter
Choose canon facts and, optionally, one saved chapter excerpt to include in the next drafting prompt.
Nothing is included automatically. Selected facts are prompt guidance, not permission to change canon. The prior-chapter excerpt is truncated if needed. Review the prepared prompt before sending it to an AI model.
Load a project to view active canon facts.
No continuity context selected.
Draft from approved outline
Choose a chapter from the currently selected approved outline. This prepares a chapter-specific request in Writer Chat; it does not generate text or modify the manuscript.
Only an approved outline can be used for this handoff.
Chapter Production Queue
Live progress derived from the selected approved outline and saved manuscript chapters.
The queue will appear when an approved outline is selected.
Queue status is calculated from saved chapter titles and manuscript approval states. It does not add a separate status record, so it cannot drift away from the manuscript. If two outline chapters share the same title, rename them distinctly for reliable matching.
Approved outlines become the approved structural reference for future context building. Editing an outline creates a new version rather than overwriting the previous version. Generated prose remains chat-only until you explicitly save it as a new manuscript draft, and remains unapproved until separately approved.
PRESERVE FIRST
Import & Continue
Paste a premise, ChatGPT/Grok excerpt, outline, notes, or manuscript. The original text is preserved before any AI action. File upload and automatic analysis are not included in this build.
SAVED SOURCES
Imports for this project
Choose a project to see its imports.
Imports are stored separately from the manuscript. Saving or viewing an import does not call the AI API. Analysis and adoption must be explicit future actions.
CANON WORKSHOP
Story Bible
Choose a project.
Build the story's reference sheet. Saving keeps fields as a draft; only the explicit approval action updates the approved snapshot. No AI request is made by this workshop.
CANON RECORDS
Characters, places, relationships, rules
Loading records…
Proposed records are not treated as approved canon. Approve individual records only after reviewing them. The current Bible snapshot and record states are separate from chat transcripts and manuscript prose.
PROJECT WORKSPACE
Writing chat
Choose a project to begin.
Messages are saved to chat history. AI generation requires a configured provider key and model allowlist.
This build saves chat messages separately from the manuscript. Generated text is not automatically approved canon and is not automatically inserted into manuscript chapters.
Provider and model router
xAI, OpenAI, and Anthropic Claude requests use authenticated server endpoints. Provider/model IDs are allowlisted, usage is recorded, and selecting a pack never triggers generation.
NEW PROJECT
Give this story a home.
This creates a saved project and an empty Story Bible. It does not call an AI API.
COMING IN A LATER BUILD
This workspace is being prepared.
This entry point is part of the foundation. Its workflow will be connected in a later build.