RoleThread Lite is a local-first app. Your datasets, sidecars, backups, settings, and local database stay on your machine unless you configure an optional backup destination.
That local-first model still needs a dependable app runtime. RoleThread needs to start cleanly, open in a familiar local app window on Windows, keep the backend bound to your own computer, shut down without leaving stray processes, and produce useful diagnostics when support is needed.
LitLaunch provides that runtime layer.
What LitLaunch Does For RoleThread
LitLaunch handles the app startup and closeout path around RoleThread:
- source launch profiles
- local app-window launch on Windows
- loopback-only local runtime configuration
- startup checks and health checks
- runtime event logging
- diagnostics reports, support artifacts, and support bundles
- runtime posture, exposure, transport, and governance reporting
- clean shutdown coordination
RoleThread stays focused on the product work: dataset editing, validation, repair, backups, cloud backup policy, preferences, and export workflows.
Source And Installed Launch
Installed Windows users normally launch RoleThread from the Start Menu or Desktop shortcut. The packaged app uses LitLaunch behind the scenes and opens a local app-style window.
Source users can run the same profile from a checkout:
python -m litlaunch.cli run --profile rolethread-webapp
Normal Streamlit browser mode is still useful for development:
streamlit run app.py
Diagnostics
When troubleshooting is needed, open Support -> Diagnostics in the app. That page is powered by LitLaunch and layered with RoleThread product context. It shows:
- runtime summary and operational snapshot
- runtime posture, exposure, transport, and governance checks
- RoleThread storage, cloud backup, data health, and support paths
- downloadable support artifacts for reports and bundles
- runtime event trail with session history
- raw runtime event trail for deeper support/debug work
Generated reports and bundles are written under .litlaunch/reports/.
Diagnostics are meant to help explain startup, profile, browser/app-window,
support, and local runtime behavior. They are support artifacts, not telemetry.
A generic redaction/privacy warning may appear; review reports and bundles
before sharing, especially if your local paths are sensitive. LitLaunch reports
posture and runtime configuration, but it does not secure Streamlit apps by
itself.
Source users and support workflows can also generate a LitLaunch diagnostics report without opening the app:
python -m litlaunch report --profile rolethread-webapp --force
Why This Matters
Earlier RoleThread prototypes proved the need for a stronger local runtime model. LitLaunch is the reusable version of that work, separated from RoleThread's dataset logic.
That split keeps RoleThread easier to maintain and makes the app a better example for other serious Streamlit projects that want a desktop-style local experience without rebuilding runtime infrastructure from scratch.