coulomb-social/docs/bubble-inventory/2026-08-12-surface-inventory.md

Bubble surface inventory — 2026-08-12

Field	Value
Workplan	CSOC-WP-0001-T01
Live product host	https://coulomb.social (Bubble)
Rebuild host	https://app.coulomb.social (Railiance)
Sources	Public HTTP, Bubble /api/1.1/meta, design extract 2026-08-09, INTENT/SEO copy
Not available this pass	Bubble editor access, Data API types (empty public meta), production DB export

1. Platform identity

Item	Value	Provenance
Bubble appname	coulomb	GET /api/1.1/meta → app_data.appname
Domain	coulomb.social	meta + DNS
App version (live)	live	meta app_data.app_version
CDN / file host	13a55be71f3a6a4c9ea356833df823c4.cdn.bubble.io	HTML + SEO image URLs
Development path	/version-test/ (robots Disallow)	robots.txt, HTTP 200
Public Data API types	empty (types: {}, no get/post)	meta 2026-08-12 — API not exposed for inventory

{
  "appname": "coulomb",
  "app_version": "live",
  "domain": "coulomb.social",
  "types": {},
  "get": [],
  "post": []
}

Implication: lossless data inventory requires Bubble editor export or an authenticated admin path — not the public Data API.

2. Product intent (marketing surface)

From Open Graph / SEO extract (docs/design-extract/coulomb-social-seo.json):

A tool to brainstorm, discuss, refine implement and promote ideas.
A social network for building stuff and influencing how stuff is built.
A marketplace for challenges, methods and solutions.
A platform to monetize curiosity.

Brand title variants seen: Research | Coulomb, Index | Coulomb, An Ideas Come To Life Community!, Coulomb Social.

3. Public / semi-public pages (HTTP probe 2026-08-12)

Path	HTTP	Title / notes
/	200	Index | Coulomb — main entry (Bubble SPA shell ~17 KB HTML)
/vw_pages	200	Research | Coulomb — primary authenticated content surface (design extract probe)
/version-test	200	Index | Coulomb — Bubble development version
/reset_pw	200	Reset password | Coulomb — Bubble built-in auth recovery
/login, /signup, /spaces, …	404	No conventional REST-style routes; Bubble uses app pages / params
/api/1.1/meta	200	App identity only; no types
/robots.txt	200	Disallows /version-test/

Bubble routes content via internal page names and pg_chunk/<id> URLs rather than clean public paths. Design extract crawled 7 routes (index + 6 pg_chunk/* pages) under an authenticated Playwright session (docs/design-extract/PROVENANCE.md).

Design-extract routes (authenticated crawl)

Route	Role (inferred)
/	Landing / index shell
/pg_chunk/1767965374100x901547861536145400	In-app content chunk (Research family)
/pg_chunk/1784305925912x322414739506593800	content chunk
/pg_chunk/1784904857383x503270614975381500	content chunk
/pg_chunk/1785501127011x363256181345746940	content chunk
/pg_chunk/1785501151364x921456713843343400	content chunk
/pg_chunk/1786106348475x662489469743530000	content chunk (+ spacing token delta)

Session probe documented as /vw_pages titled Research | Coulomb.

Gap: human-readable Bubble page names (editor “Pages” list) and full navigation graph are not in the extract artefacts — only token diffs per chunk URL. Next inventory pass should export page list from Bubble editor.

4. UI surfaces and components (from design extract)

Signal	Observation	Rebuild mapping
Primary surface	Research / pages workspace	Space detail + markdown pages (ADR-0002)
CTA verbs	content, favorite, format, filter, add, note	Feature candidates for cut (T03)
Material icons	content_copy, favorite_border, filter_list, add_circle_outline, note_alt, …	App chrome icons
Forms	10 textareas, 6 inputs in crawl sample	Authoring / filters
Stack signal	Bubble (+ webflow heuristic noise)	Rebuild is Django, not Bubble
Client libs in page	marked, highlight.js, clipboard.min.js	Markdown-oriented UX already present in Bubble

Visual system is fully captured under docs/design-extract/ (tokens, DESIGN.md, preview). Do not re-extract for inventory unless pages change materially.

5. Content model hypothesis (spaces / pages → rebuild)

Prefer spaces / pages / markdown-shaped content for migration (workplan guidance + ADR-0002).

Bubble concept (hypothesized)	Evidence	Rebuild target
Research / vw_pages workspace	Title, extract probe	Space + pages/*.md in Forgejo
Page / chunk (pg_chunk ids)	6 chunk URLs in extract	One markdown file per page slug under content_root
Rich text / markdown rendering	marked + highlight.js on live HTML	CommonMark + sanitized HTML in app
Favorites / likes	CTA verb counts	Later social feature; not required for content MVP
Filters / lists	filter CTA	Space list filters later
Binary assets	Bubble CDN file URLs	assets/ in Forgejo repo or object store later
Users / members	Bubble account system + reset_pw	NetKingdom OIDC Members (done path)
Tenancy	Single Coulomb community today	tenant:coulomb instance

Confidence: medium for “Research pages ≈ space content”; low for full data-type graph until editor export.

6. Workflows (partial)

Workflow area	Status	Notes
Member signup / login	Live on Bubble	Rebuild uses NetKingdom; Case B residual CSOC-IN-0001
Password reset	/reset_pw	Platform-owned on rebuild
Page view / research	/vw_pages	Core product path
Content create/edit	Inferred from textareas + content CTAs	Rebuild: edit in Forgejo (T05 path)
Marketplace matching	Marketing claim only	Feature-cut candidate (T03)
Monetization	Marketing claim only	Out of v1 unless T03 expands

Backend workflows (API workflows, backend triggers) are not inventoriable without Bubble editor or API workflow listing.

7. Integrations and external services

Service	Role	Rebuild ownership
Bubble.io hosting	App runtime, DB, CDN	Exit target
cdn.bubble.io	Images / static files	Migrate assets off CDN at cutover
Cloudflare (on Bubble CDN path)	Image resize (cdn-cgi/image)	Not required on rebuild
highlight.js (cdnjs)	Code block styling	Optional in app CSS
marked (jsDelivr)	Client markdown	Server-side markdown in Django instead
clipboard.js (Bubble CDN file)	Copy UX	Optional
(unknown) email provider	Bubble account mail	NetKingdom mail path for registration
(unknown) payment	Monetization claim	Not observed publicly

8. Data types (blocked)

Public meta returned no types. Expected classes (to confirm on export):

• User / User profile
• Space / Project / Group (name TBD)
• Page / Chunk / Document
• Favorite / Follow / Membership
• Challenge / Method / Solution (marketplace — optional)
• Message / Comment (if present)
• File / Image references

T02 owns dump + mapping once export is available.

9. Inventory completeness

Area	Completeness	Next action
Public hosts & meta	high	refresh if domain changes
Visual design system	high	already in docs/design-extract/
Named Bubble pages list	low	editor screenshot/export
Workflows	low	editor Backend workflows list
Data types & fields	none public	Bubble CSV/JSON export or API key
Content corpus size	unknown	export pages → markdown rehearsal
Member count / PII volume	unknown	export under privacy rules

10. Priority order for migration (recommended)

Aligned with stage-1 cut (docs/capability/stage-1-cutover.md):

1. Members → NetKingdom JIT / assisted transfer (no Bubble passwords)
2. Spaces + pages (Title, Abstract, Visual + markdown body) → app metadata + Forgejo
3. Page copy / transfer support data as needed
4. Memberships beyond tenant-wide defaults
5. Favorites / marketplace / monetization → later (not stage 1)
6. Bubble workflows → rewrite only if product need returns (out for parity)

Related

• Capability map (live product): docs/capability/existing-bubble-capability-map.md
• Parallel hosts: docs/deploy.md
• Content model: docs/adr/ADR-0002-space-content-forgejo-markdown.md
• Auth extract: docs/design-extract-auth.md
• Workplan: workplans/CSOC-WP-0001-bubble-io-exit-assessment.md