Jay Griffin

What Is It?

A personal tracker for consuming your favorite creators' work. Add a creator, ingest their canon, curate what you want, track your progress. Gamified completionism for intentional consumption.

Core loop: Browse creators → see % complete → pick something to read/watch → mark done → watch the number go up.

Core Concepts

Concept	Description
Creator	A person or entity whose work you follow (blogger, YouTuber, podcaster, docs maintainer)
Work	A single piece of content (article, video, episode, doc page)
Canon	The full set of works by a creator — discovered via sitemap, RSS, scraping, or manual entry
Queue	Works you've approved/curated from the catalog — the ones you actually want to consume
Progress	% of queued works marked done, per creator

Data Model

`creators` table

Column	Type	Purpose
`id`	TEXT PK	UUID
`name`	TEXT NOT NULL	Display name ("Kent C. Dodds")
`url`	TEXT NOT NULL UNIQUE	Homepage URL
`description`	TEXT	Why you follow them
`image_url`	TEXT	Headshot or avatar
`type`	TEXT DEFAULT 'author'	`author` / `docs` / `youtube` / `podcast` / `newsletter`
`created_at`	TEXT	When added

`works` table

Column	Type	Purpose
`id`	TEXT PK	UUID
`creator_id`	TEXT FK	References creators.id
`url`	TEXT NOT NULL UNIQUE	Direct link to the work
`title`	TEXT NOT NULL	Title of the work
`description`	TEXT	Subtitle, excerpt, or scraped summary
`status`	TEXT DEFAULT 'candidate'	`candidate` → `queued` → `done` → `dropped`
`difficulty`	TEXT	`light` / `medium` / `dense` (optional, user-set)
`topics`	TEXT (JSON)	Tags like `["react", "testing", "patterns"]`
`note`	TEXT	Your personal annotation
`rating`	INTEGER	1–5, set after reading
`read_count`	INTEGER DEFAULT 0	Times consumed (re-reads/rewatches count)
`discovered_at`	TEXT	When ingested from catalog
`queued_at`	TEXT	When approved into queue
`last_read_at`	TEXT	Last consumption timestamp

Indexes

idx_works_creator on works(creator_id)
idx_works_status on works(status)

Computed Stats (queries, not stored)

Canon % = COUNT(status='done') / COUNT(status IN ('queued','done')) per creator
Total works = all works for a creator
Queued = approved but not yet consumed
Re-read favorites = ORDER BY read_count DESC, rating DESC
Latest unread = next queued work by discovered_at

Ingestion Strategy (priority order)

When you paste a creator's URL, try these in order:

RSS feed — auto-discover via <link rel="alternate"> in HTML head. Richest data: title, URL, description, date.
Sitemap — fetch /sitemap.xml. Gives all URLs, sometimes with lastmod dates. No titles (derive from URL path).
llms.txt — fetch /llms.txt. Newer convention, curated page list. Growing adoption.
Scrape links — fetch the page, extract all internal <a> hrefs, dedupe, filter obvious non-content (assets, auth, etc.).

All ingested works start as candidate. Approval UI lets you bulk-select which ones to queue.

Pages & Views

`/canon` — Main page

The hero view. Grid of creator cards.

Each card shows:

Creator image/avatar
Name + description
Progress ring (circular % indicator)
Stats line: "34/51 read · 3 favorites · 14 queued"
"Next up" — title of the next queued work, clickable
"Random" button — pick a random queued work
"Top picks" — 2-3 most re-read or highest-rated works

Sorted by: most recently active, or lowest % (motivational), user choice.

`/canon/[creatorId]` — Creator detail

Full catalog view for one creator:

Creator header (image, name, description, big progress ring, stats)
Filter bar: status (queued/done/candidate/all), difficulty, topic tags
Works list: cards or compact rows, each with title, status badge, difficulty, rating stars, read count
Bulk actions: "Queue all candidates", "Mark selected as done"

`/canon/discover` — Add a creator

Paste a URL
App tries RSS → sitemap → llms.txt → scrape
Shows results: "Found 47 articles from kentcdodds.com"
Creator name/description fields (pre-filled if possible)
Approval grid: checkboxes on each discovered work
"Queue selected" / "Queue all" buttons

Route Structure

src/app/canon/
  page.tsx                    # Main grid of creator cards
  canon.module.css            # Styles for main page
  [creatorId]/
    page.tsx                  # Creator detail / catalog view
  discover/
    page.tsx                  # Add new creator + ingest

Server Actions

addCreator(name, url, description, imageUrl, type) — insert into creators
ingestWorks(creatorId, works[]) — bulk insert discovered works as candidates
updateWorkStatus(workId, status) — queue, done, drop
markRead(workId) — increment read_count, set last_read_at, set status=done if first read
rateWork(workId, rating) — set rating
bulkQueueWorks(workIds[]) — move candidates to queued

Build Order

Tables — create creators + works in Turso
Data layer — src/lib/reading.ts with types + query functions
Server actions — src/actions/reading.ts
Discover page — paste URL, ingest, approve works
Main page — creator cards with progress rings
Creator detail page — full catalog, filters, mark done
Sidebar link — add to nav
Polish — random button, re-read tracking, stats

Future Ideas (not building now)

YouTube/podcast support (different embed/player)
Spotify Wrapped-style "year in review" stats
Public profile: "biggest fan" leaderboard per creator
Import from Pocket/Instapaper/browser bookmarks
Browser extension: "Add to Canon%" from any page
AI summaries of articles you've read
Spaced repetition: resurface articles you rated highly after N days