The TimelineAI journey
A transparent view from the first platform scaffold to the next generation of cited, reviewable AI storytelling. Shipped means demonstrated in code; planned means scoped and sequenced, not date-promised.
Built so far
The August 2026 MVP and current reliability work
Platform contract, schema, auth & CRUD
TimelineAI started as a first-class platform app with shared Prisma persistence, Hub SSO, server-derived ownership, authorization rules, and timeline/event APIs.
Accessible editor & live preview
A drag-and-drop and keyboard-operable editor paired content entry with an immediate visual preview and client/server schema validation.
Dashboard, ownership & self-publishing
Signed-in creators gained a dashboard, edit flow, private drafts, and direct publishing under the platform's shared identity.
MVP accessibility & end-to-end evidence
Automated accessibility checks and real guest, authenticated, and admin journeys completed the twelve-phase MVP with documented architectural decisions.
Visual director UI
Compare 2-3 accessible visual directions side by side as live mini-previews with plain-English rationale and the heuristic's own inputs, apply one in a click, and undo it in a click.
Where the story goes next
The cited AI storytelling studio epic (#298)
Version history & collaborative review
Immutable versions, structured diffs, anchored comments, reviewer capabilities, and reversible restore give both human and AI-authored changes a safe publishing workflow.
Provenance-aware media enrichment
Suggest relevant allow-listed media with creator, license, attribution, and editable grounded alt text—never an untraceable image scrape.
Reviewed multilingual timelines
Create locale variants through glossary-aware translation proposals that preserve names, dates, citations, and links, with RTL and long-text layout verification.
Narrated story mode & Vionto handoff
Turn a reviewed timeline into an editable narration, accessible playback, captions, transcript, and a versioned scene handoff to Vionto.
Branching scenarios & what-if comparison
Fork a factual baseline into named alternatives, compare outcomes, and propose risks or decision points while making assumptions visually impossible to confuse with history.
Living roadmaps from GitHub & TasksAI
Read selected issues, PRs, milestones, and TasksAI work into idempotent proposed updates with drift/conflict review and a cited “what changed” digest.
Ask & explore with cited answers
Let visitors ask what changed, where the gaps are, and which events support a conclusion; prefer deterministic calculations and cite every factual answer.