Appearance
How a video gets published
needs verificationFollow one Publication, a three-chapter research video. The capture and core workflow below are implemented. The final production handoff to the content management system (CMS), and minting a digital object identifier (DOI), remain incomplete or unverified.
Here is the journey to keep in mind: a researcher explains a paper, the recordings become one video, the researcher confirms the result, and the public page makes the research easy to watch and cite. A coordinator can help create the work, but the subject researcher provides approval.
See a finished Publication Plus on lt.org to picture the destination. That example was produced editorially; the mobile Publication journey below is a different route to a research video.
Capture a draft
The researcher taps the center button, selects Publication and creates an LT Work draft (draft_new). They select one to five OpenAlex source papers, then record Question, Method and Findings. A coordinator starts by selecting the subject researcher. The app provides a teleprompter, front-camera face alignment, preview and cover-photo capture.
Keeping a take starts multipart upload control through lt-core; the video parts travel directly to object storage. A work remains incomplete until its required source papers, cover and chapters are present. The Open Researcher and Contributor ID (ORCID) proof gate is role-specific: current Publication and Profile Video configuration requires the creator's linked ORCID for researcher-created work, but not coordinator-created work. This differs from the brief's universal-proof wording.
Submit, process and confirm
Submitting checks readiness and changes the work to Submitted. A coordinator-created work waits for the subject's approval (submitted_pending_researcher); a draft subject receives an emailed link. Once processing can start, lt-core saves a request snapshot and queues the Media Processing Engine (MPE) job (submitted_processing).
MPE returns renditions, subtitles and a poster. lt-core validates the callback, registers the existing files in Cloudflare R2 and projects a Video. A failure produces submitted_processing_failed; a revision request can return the work to a draft. Successful processing automatically supplies the review approval marker because human editorial review is disabled in the current minimum viable product (MVP). A draft subject must first register, resolve the draft into a real researcher and have an accepted coordinator connection before final confirmation. The researcher then confirms (approved_confirmed) or declines back to revision.
Publish and reach the public site
LTWorkPublishService requires a confirmed work, real researcher subject and current ready Video. It freezes attribution (names, ORCID and OpenAlex identifier), timestamps publication, transfers ownership to the published researcher and projects the website payload. The Filament Run mock Publish action calls this service and is visible for approved_confirmed works. Its visibility code has no environment restriction, despite the action's name. Public publish, unpublish and republish API endpoints remain placeholders. The production automation and who should use this action need team confirmation.
Core exposes website payload read routes authenticated for Model Context Protocol (MCP) integrations. A route does not establish that Strapi consumes it today. Hugo's build-time Strapi reads are implemented. The website embeds player configuration in its HTML; it does not directly call the authenticated core Video API as the brief suggests.
What is still missing
The target is an immutable, citable LT video with its own DOI assigned at publication. Current core API responses return doi: null and public_url: null; the embed snippet field was removed. The website payload's DOI currently comes from the source paper, not a minted video DOI. Do not treat it as evidence that the video has been registered.
Published videos are retained; a new processing output supersedes the previous Video instead of deleting it. Profile Video follows the capture flow without source papers or DOI eligibility. Publication Plus is editorially produced with five landscape chapters and one paper; its core website projection is not implemented.
For maintainers: source references
These references require repository access. Everything needed to understand this page is explained above.