Roadmap

From private desktop preview to a broader workflow platform.

LiraLora is being built around guided AI workflows where continuity, memory, and replay matter more than one-shot prompting.

The first foundation is a local-first desktop app for managing longer AI runs, reviewing memory, keeping work organized by project, and reusing successful workflow history.

This roadmap separates private-preview work from public release polish and longer-term workflow expansion.

In private preview: run control, reviewable memory, project history, image workflows, early music drafting, custom workflows, and coding-assistant workflows

Before public release: installers, signing, updates, QA, safer credentials, and account readiness

Later: workflow packs, sharing, collaboration, real audio/video providers, and selective cloud offload

Explore current workflows See how it works

Product arc

The roadmap widens in phases so the product can stay coherent while the platform grows.

Now / private preview

The local-first desktop foundation is being tested privately

Current private-preview work focuses on the core workflow experience: starting long-running AI tasks, tracking their progress, reviewing memory, organizing work by project, and reusing successful runs across images, music drafts, custom workflows, and coding-assistant flows.

Track active and recent AI runs, including background work, cancellation, provider/model selection, and related memory
Keep workflow history organized by project so prior runs can be reviewed and reused
Review memory before it becomes durable, with controls for search, manual creation, archive, restore, and deletion
Plan and generate images using project context, provider settings, and reusable memory
Draft music ideas with editable structure, prompt/export support, and local preview where supported
Create and run built-in or custom multi-step workflows
Use local models and OpenAI-compatible endpoints through private provider settings
Explore coding-assistant workflows through a local CLI integration in private builds

Release polish

Prepare the desktop app for a public release

The next milestone is making the private build safe and reliable enough for broader use. That means packaging, signing, updates, clearer release channels, stronger QA, safer local credential handling, and account surfaces that match what the product actually exposes.

Public installers and platform packaging for macOS, Windows, and Linux
Code signing, auto-update, release notes, and clear update channels
Safer handling for local provider credentials before hardened releases
Account, checkout, billing, and entitlement flows only where they match real product surfaces
End-to-end QA for local providers, API-backed memory, project access, and provider configuration
Broader provider coverage without implying managed hosted usage is complete before it is ready

Workflow expansion

Expand into reusable workflow families

Once the desktop foundation is stable, LiraLora can grow by adding coherent workflow families rather than isolated tools. New areas should build on the same memory, history, run-control, provider choice, checkpoints, and replay model.

Richer replay, branching, comparison, and continuation flows
Workflow templates, packs, sharing, and team libraries
Real audio-generation providers and stronger artifact handling
Team and collaboration features built on clear project access boundaries
Selective cloud offload and training/fine-tuning support for work that is impractical to run locally

Workflow horizons

These are the kinds of workflow families that fit the longer-term shape of the product.

Future workflow family

Coding and app-building workflows

Coding workflows are a natural fit for LiraLora’s memory and replay model. Private-preview work explores how project context, durable run history, and local coding-agent integrations can support ask, analyze, and edit-style workflows without replacing dedicated coding tools.

Creative media

Music generation pipelines

Music work starts with planning: structure, prompts, editable drafts, and reusable ideas. Future provider work can connect those plans to real audio generation, richer artifact handling, and iteration over completed tracks.

Continuity-heavy media

Video generation and shot continuity

Video raises the stakes for continuity. Characters, scenes, story beats, edits, and revisions all become easier to manage when the system can remember what mattered and continue from the right stage.

Reusable asset systems

Asset pack generation

Recurring items, environments, props, style variants, and visual identity systems are a strong fit for workflow-guided generation with approvals, lineage, and shared memory across many outputs.

Educational expansion

Advanced guided learning materials

The educational side can grow from early-reader concepts into lesson systems, subject-specific material, larger textbook-like structures, and multi-level content that remains controlled, reusable, and easier to revise.

Learning experiences

Children’s learning experiences

Illustrated readers, activity sets, workbook companions, and adaptive learning materials benefit from continuity when recurring characters, style consistency, reading level, and content review all matter at once.

Roadmap principles

The expansion story only works if the product keeps feeling like a guided system with a clear center of gravity.

Private-preview work and public availability are labeled separately.
Drafting and planning tools are not presented as completed generation providers.
New workflow families should reuse the same memory, history, run-control, replay, and provider foundation.
LiraLora stays local-first where practical, with cloud/offload paths added only when they clearly improve the job.

Understand the foundation

How it works Explore workflows Memory vs context FAQ

The roadmap makes the most sense alongside the pages that explain how workflows, memory, and provider choices fit together.