Founder · Product Architect · Software Engineer
In development · Testing
Instead of publishing another article every time a report changes, F-PUB brings related reports into one evolving Story — showing readers what is known, where it came from and what changed.
Overview
Sports news moves quickly, but readers still need to know what is actually confirmed, who reported it and how the picture changed over time.
F-PUB treats a real-world event as one evolving Story. New reports, official statements, denials and meaningful updates are connected to that same Story instead of becoming separate articles that force the reader to reconstruct the timeline.
The current product includes a public Hebrew sports site, story and entity pages, search, methodology and corrections, alongside an internal newsroom for sources, editorial review and operational work.
Product scope
The public experience and the newsroom are designed around the same underlying story, source and evidence model.
A single story can collect new reports, official updates, disagreements and corrections while keeping one stable destination for the reader.
Stories show the publications and reporters behind the information, with clear links back to original reporting.
Story states and summaries help distinguish reporting, independent confirmation, official information and unresolved uncertainty.
Meaningful developments remain connected to the same story so readers can understand how the information evolved.
Team, league, player and match surfaces connect news with structured football context instead of separating data from editorial content.
Readers can move between stories, teams, players, leagues, reporters and publishers rather than navigating only a chronological feed.
Public methodology and correction history make editorial changes visible instead of silently replacing the previous version.
An internal workspace brings review queues, sources, story editing, corrections and operational signals into one editorial console.
AI in the product
The implemented AI workflows help organize and interpret reports rather than simply rewrite articles. F-PUB remains in development and testing. AI is not a journalistic source, and sensitive publication decisions retain human editorial review.
Classify whether an incoming report belongs to the publication and identify its sporting context before it moves further through the workflow.
Assess whether a new report belongs to an existing Story, reducing disconnected versions of the same developing event.
Turn reports into structured claims while retaining source references and the certainty expressed by the source.
Analyze whether reports are independent, derivative or attributed to another source, preserving context behind repeated coverage.
Generate editorial drafts from the available evidence and story context, with validation rather than treating model output as automatic publication.
Assess the significance of updates and validate structured outputs so review can focus on changes that matter to the Story.
Users & surfaces
The product serves readers directly while giving the editorial side a separate environment for review and publication work.
Gets one current story with visible sources, status, updates and sports context instead of chasing the same event across multiple articles.
Works from a dedicated newsroom that surfaces items needing attention, supports review and keeps updates tied to the story history.
Publishers, reporters and official sources are represented as part of the story context so attribution remains visible to the reader.
Core workflow
Automation reduces repetitive work, but publication remains grounded in sources and editorial control.
Architecture
Public reading, newsroom operations, ingestion and intelligence share a common story and evidence model while remaining separate product layers.
The architecture keeps editorial history and source context reusable across the public site and newsroom instead of rebuilding the same event in separate representations.
Engineering decisions
Selected decisions show how editorial principles become durable product and data rules.
Related reports update one canonical Story rather than producing a new page for every source or development.
This changes the underlying data model: sources, claims, revisions and sports context attach to the Story, allowing the public page to evolve without losing the history of the event.
Where information came from is stored as part of the story rather than added later as editorial decoration.
The system keeps source attribution, timing and evidence relationships connected to the information they support, which makes it possible to explain independent confirmation and link readers back to original reporting.
When a source or story changes, the product can preserve the change instead of treating the latest text as the only truth.
Versioned source and story records support meaningful updates, corrections and editorial review while keeping a stable public Story URL.
AI helps classify, connect, summarize and validate information, but publication stays grounded in evidence and human review.
AI outputs are structured and checked against the story context before they can support editorial work. The product is designed so uncertainty can remain visible instead of being converted into confident copy.
A material correction is not just a text edit — it has a visible place in the story history and a public corrections surface.
This makes accountability a product behavior: readers can see that something changed, while the editorial side keeps the correction connected to the affected story and revision.
Scores, fixtures and standings should disappear or degrade gracefully when the underlying data is not current enough.
Sports data enters through an adapter layer with explicit freshness semantics, allowing public surfaces to avoid presenting stale structured data as live information.
Engineering evidence
The repository combines public publishing, editorial operations and multiple data pipelines, with automated tests across domain rules and integrations.
Current codebase snapshot
Technology
The technology is grouped by the responsibility it serves in the publishing and newsroom system.
Product principle
F-PUB is designed to show the difference between a report, independent confirmation, official information and a correction. The product adds value by organizing the evidence around a story rather than pretending every new headline is a new fact.
Next stage
F-PUB is in development and testing as we prepare its next stage. We’re preparing to open conversations with strategic investors who understand sports, media and AI-assisted editorial technology.