Curiosity Atlas v1.0.0 — App Review Summary

Loaded Source URL

LOADED_SRC: https://producingtechnology.com/65-apps/wuirene_184763_15200472_json.html

App Overview

Curiosity Atlas v1.0.0 is a personal knowledge-mapping app that visualizes a network of ideas as a connected graph. The aesthetic is moody and editorial: a deep black canvas, a thin neon-green hairline at the top, and gold serif accents on the title and toggle buttons. Each item in the graph belongs to one of five typed categories — concept, book, thinker, question, and field — each with its own color and icon. A header strip surfaces three running totals: 142 Explored, 89 Connections, and 17 Open Qs. A daily prompt sits at the bottom of the page: Pick one open question and write three sentences you couldn't have written last week.

Behavior Summary

The interface is split into a left rail (search + node list) and a main canvas that toggles between two views:

Header / Stats bar: Shows app name and version, a Graph/List view toggle, and three counters (Explored / Connections / Open Qs).
Left rail: A search box, an Add button, a row of type-filter pills (concept, book, thinker, question, field), and a vertical list of every node with its title and type label.
Graph view (default): Renders each node as a colored circle with an icon, laid out across the canvas with thin dashed lines connecting related nodes. The sample shows five nodes — Emergence (concept), Mycology (field), Gödel, Escher, Bach (book), Ada Lovelace (thinker), Strange Loops (concept), and a question node Can a system fully understand itself?
Node detail panel: Clicking a node (e.g., Ada Lovelace) opens a right-side panel with the type label, title, key dates (1815–1852), a "mastery" level (familiar), the date added (2026-01-20), a one-line description, topical tag pills (computing, mathematics, history), and a Connections list showing typed edges (e.g., "← references Gödel, Escher, Bach"). The corresponding edge in the graph also gets emphasized with a label.
List view: Toggling List swaps the canvas for a card grid. Each card shows type, title, a short description (e.g., for the book: Douglas Hofstadter · 1979; for Mycology: a description of fungi and forest networks), a status pill (read, exploring, curious, familiar), and the same topical tag pills.
Daily prompt + date: A footer line shows the journaling prompt on the left and the current date (2026-02-25) on the right.

Things That Didn't Work As Expected

Header counters don't match what's visible: The header advertises 142 Explored, 89 Connections, and 17 Open Qs, but the graph and list each only show 6 nodes and a couple of visible edges. Either the numbers should reflect the actual loaded dataset, or they should be labeled as aspirational/lifetime stats — right now they read as misleading.
Graph layout feels static and sparse: The five nodes are spread far apart with very faint dashed connections, making relationships hard to read at a glance. There's no visible force-directed physics, no zoom/pan affordance, and no clustering. The empty black canvas dominates the screen.
Edges are hard to see: Connection lines are extremely low-contrast dashed strokes against pure black. Until a node is selected (which highlights one edge with a "references" label), it's almost impossible to tell which nodes are actually connected.
Type filters appear non-interactive: The five colored pills (concept, book, thinker, question, field) look like filters but it's not clear whether clicking them actually filters the graph or list — there's no visible active state and no count of matched nodes.
Add flow is opaque: The Add button is prominent but the affordance for what it does (open a modal? inline form?) isn't communicated. There's no indication of whether new nodes get edges drawn automatically or require manual linking.
Search lacks feedback: A "Search nodes…" box sits at the top of the rail but there's no live filtering or result counter visible during typing, and no empty-state message for zero matches.
Node detail panel is bare: The Ada Lovelace panel shows a one-line summary, a few tags, and one connection, but there's no notes field, no journal entries, no "last revisited" timestamp, and no way to add a new connection from inside the panel — features you'd expect from a knowledge tool that tracks "mastery" over time.
"Mastery" levels lack a legend: Status chips (exploring, read, curious, familiar) appear on cards and in the detail panel, but there's no key explaining the progression or how mastery is updated.
Daily prompt is decorative: The footer line about writing three sentences is a great hook, but it isn't actionable — there's no button to start journaling, no link to the related question node, and no record of past prompts.
Type icons are tiny: The diamond/book/brain/question/leaf icons inside each node circle are small and a bit illegible on the graph view, especially the question mark which can be confused with a generic placeholder.
No dataset import/export: For an app that's clearly built around a JSON backbone, there's no obvious way to import your own JSON, export the current atlas, or share a snapshot.

Suggested Improvement Prompt

Improve the Curiosity Atlas single-file web app so the graph feels alive and the knowledge tool actually supports the daily-practice ritual it advertises. Specifically:

Make the graph interactive: Add a force-directed layout (e.g., d3-force or cytoscape) with drag, zoom, and pan. Cluster nodes by type or by tag, and animate transitions when filters or selections change.

Improve edge legibility: Replace the low-contrast dashed lines with solid strokes that adopt the source node's color at lower opacity, and label the edge type ("references," "inspired by," "contradicts") on hover, not just on selection.

Wire up the type filters: Clicking a type pill should filter both the graph and the list view to matching nodes, with a clear active state, a count badge, and a multi-select mode (shift-click) so users can combine filters.

Make stats reflect loaded data: Compute Explored / Connections / Open Qs from the current JSON. If the larger numbers are intentional lifetime totals, label them ("Lifetime") and add a smaller "in view" pair beside them.

Build a real Add flow: Open a modal that takes title, type, tags, mastery, description, and an optional list of connections (multi-select from existing nodes), and animate the new node into the graph with edges drawn.

Live search with highlighting: Filter both the rail list and the graph as the user types; dim non-matching nodes in the canvas, show a "no results" empty state, and support fuzzy matching across title, tags, and description.

Expand the node detail panel: Add a free-form notes field, a journal log with timestamped entries, an "update mastery" control, and an inline "Add connection" picker. Show the date last revisited and a small spark line of journaling activity.

Mastery legend & progression: Add a small legend (curious → exploring → familiar → fluent) with hover tooltips, and let users update mastery from any card or the detail panel — logging the change to the node's history.

Make the daily prompt actionable: Tapping the footer prompt should pick a random open question, open its detail panel, scroll the journal field into view, and timestamp the entry on save. Keep a rotating archive of past prompts.

Import / export JSON: Add explicit "Load JSON" / "Download JSON" controls so users can bring their own atlas, share snapshots, and back up their data — true to the app's JSON-driven framing.

Polish the icon set: Use slightly larger, higher-contrast SVG icons inside each node circle, with consistent stroke weight, so concept/book/thinker/question/field are instantly distinguishable on the graph.

Optional light theme: Offer a light variant for daytime reading; the gold serif title should adapt rather than disappear against a pale background.