lt-core

needs verification

Purpose

lt-core is the backend and system of record for accounts, researcher relationships and the LT Work workflow. It owns lifecycle rules and application programming interface (API) contracts, while Strapi owns editorial content management and Hugo owns public presentation.

Start with lt-core staging when the team gives you access. Think of it as the team's working records: who the researchers are, who is helping them, and where each video is in its journey.

What it does

Your first visit

The left menu groups people, LT Works, videos and affiliations. Researchers holds scientific profiles; Researcher Drafts holds invited people who have not registered. LT Works shows work in progress and its status. Videos holds playable outputs. You do not need to learn the underlying data model before exploring these screens.

Local development dashboard, captured 10 September 2026 after signing in at local lt-core. Counts are sample local data, not production figures.

Open LT Works from the left menu to see each video's title, type and progress. This is the same local lt-core development instance; the screenshot contains sample records.

The records behind the screens

Domain	What the records mean
User, Researcher, Researcher draft	A sign-in account, its scientific profile, or a person invited before registration
Coordinator, Coordinator group, Connection	An individual or shared working identity and its relationship with a researcher
Institution, Project	Organizations and project associations used in profiles
LT Work, LT Work type, Chapter	The workflow record, its format rules and individual recorded segments
Processing run, Video	A saved processing attempt and its playable output
OpenAlex work, DOI reference, File upload	A source paper, registration scaffolding for a Digital Object Identifier, and stored-file metadata

Core handles email one-time codes, password and Open Researcher and Contributor ID (ORCID) login; profile and coordinator relationships; OpenAlex-assisted lookup; multipart upload control; submission and approval; Media Processing Engine (MPE) results; and website payload projection. The Filament admin panel provides inspection and management. The Model Context Protocol (MCP) server serves agent integrations.

For maintainers: technology

Laravel 12 and PHP 8.2+, PostgreSQL 16, Redis, Sanctum bearer-token authentication, Filament 4, Spatie permissions and Laravel MCP. Version requirements come from composer.json; PostgreSQL is documented in AGENTS.md.

How it connects

The mobile app sends workflow requests and uploads media directly to object storage using Core's upload control plane. A saved request goes to MPE; its callback returns processed-media metadata. Core validates existing R2 objects and projects Video records. Website payloads are available through machine-authenticated internal routes.

The Video configuration API is authenticated and scoped to researcher/coordinator work visibility. It is not an anonymous public website endpoint.

Who can do what

“Partial” means a restricted scope or a separate authentication surface. Administrator here means the Core role; Editor belongs to Strapi.

Action	Anonymous	Researcher	Coordinator	Administrator	Editor
Sign in or register through public auth routes	Yes	Yes	Yes	Partial	No
Use mobile-domain API	No	Yes	Yes	Partial, impersonation	No
Manage own researcher profile	No	Yes	No	Yes, admin panel	No
Manage coordinator groups and drafts	No	No	Yes	Yes, admin panel	No
Create and submit LT Works	No	Yes, own subject	Yes, permitted subject	Partial, impersonation	No
Approve subject work and confirm final result	Partial, draft email link only	Yes, subject only	No	Partial, impersonation	No
Inspect all users, works and processing runs	No	No	No	Yes	No
Run mock Publish after final confirmation	No	No	No	Yes	No

Environments and links

Localhttp://core.lt.localhost

Staginghttps://core.stg.lt.org

Productionhttps://core.lt.org

Environment base URLs are documented in staging infrastructure and production cutover evidence. These are documented destinations, not an availability test. Ask the team lead for a staging admin account; credentials do not belong in this site.

Where to read more

People and roles, what an LT Work means and the publication journey.

Known gaps

• DOI registration is scaffolding. The LT Work API returns null DOI and public URL; the website payload uses the source paper DOI. See DOI.
• Publish, unpublish and republish API routes are placeholders. Filament's Run mock Publish does call the publish service after researcher confirmation. Production automation remains unverified.
• Human editorial review is disabled in the minimum viable product (MVP). Successful MPE result handling sets the approval marker.
• Publication Plus website projection is unsupported. Automatic CMS delivery and page backreferences still need verification.
• The domain skill's February note saying no LT Work state machine was found is stale. Current minor-status enum and model take precedence.

All unresolved items are tracked in Open questions.

For maintainers: source references

These references require repository access. Everything needed to understand this page is explained above.

• repository boundaries
• domain models
• type configuration
• API routes
• authentication routes
• MCP routes
• composer.json
• AGENTS.md
• MPE contract
• website payload routes
• Video routes
• role seeds
• mobile restriction
• LT Work routes
• final confirmation checks
• Filament actions
• staging infrastructure
• production cutover evidence
• domain skill
• ORCID identity model
• developer actions
• result service
• minor-status enum