diff --git a/README.md b/README.md
index e7922fc..4d4fba6 100644
--- a/README.md
+++ b/README.md
@@ -1,58 +1,13 @@
 # Malfestio
 
-Malfestio is a learning OS: flashcards + notes + lectures + articles, designed for daily study.
+Malfestio is a social learning platform: flashcards + notes + lectures + articles, designed for daily study.
 
 Social layer: publish/share/remix learning artifacts; follow curators; discuss.
 
-## Personas
+## Documentation
 
-- **Learner**: studies daily; imports content; wants fast "review queue".
-- **Creator**: makes decks/notes; publishes updates; wants feedback + forks.
-- **Curator/Teacher**: bundles content into learning paths; annotates lectures/articles.
-- **Moderator/Community admin**: handles reports, takedowns, spam.
-
-## Principles
-
-- Local-first study experience; offline study must not feel "second-class".
-- Shareable artifacts are portable: Lexicon-defined schemas + stable IDs.
-- Privacy by design: progress + recall history are private unless explicitly
-  shared.
-
-### Data Model
-
-- Note: markdown + structure + citations + links to sources.
-- Card: front/back (+ optional cloze, audio, image, code block).
-- Deck: ordered/clustered cards (+ metadata, tags).
-- Lecture: external URL + outline + timestamps + linked notes/cards.
-- Article: URL + extracted text (readability style heuristics) + highlights +
-  linked notes/cards.
-- Collection/Path: curated bundle of decks + notes + sources.
-
-## System Architecture
-
-### Frontend (SolidJS)
-
-- App shell + router-driven workspaces (Library / Study / Create / Social).
-- Signals as primary state primitive; keep study session state in signals/store.
-
-### Backend (Rust)
-
-- Axum API gateway: REST/XRPC-ish endpoints, tower middleware, typed extractors.
-- Services (logical, not necessarily microservices):
-    - Identity/Auth service (local + optional ATProto OAuth integration)
-    - Content service (notes/cards/decks/sources)
-    - Study service (queue generation + grading + scheduling)
-    - Social service (follows, feeds, comments, notifications)
-    - Search service (indexing + query)
-    - Moderation service (reports, takedowns, rules)
-
-### Storage
-
-- Postgres: canonical app DB (users, private study state, cache of published records).
-- Object storage: images/audio, extracted article snapshots (if you store them).
-- Search index: separate system (Meilisearch/Typesense/ZincSearch-pick one later).
-
-### Eventing
-
-- Internal outbox pattern (DB table) for:
-    - reindex jobs, notification fanout, federation publish steps
+- [Personas & Principles](./docs/personas.md) – Target users and design philosophy
+- [Architecture](./docs/architecture.md) – System components and data model
+- [Information Architecture](./docs/information-architecture.md) – Navigation and URL structure
+- [Data Model Mapping](./docs/data-model-mapping.md) – Lexicon to database mapping
+- [Roadmap](./docs/todo.md) – Development milestones
diff --git a/docs/architecture.md b/docs/architecture.md
new file mode 100644
index 0000000..2664fa0
--- /dev/null
+++ b/docs/architecture.md
@@ -0,0 +1,43 @@
+# System Architecture
+
+This document describes the technical architecture of Malfestio.
+
+## Frontend (SolidJS)
+
+- App shell + router-driven workspaces (Library / Study / Create / Social).
+- Signals as primary state primitive; keep study session state in signals/store.
+
+## Backend (Rust)
+
+- Axum API gateway: REST/XRPC-ish endpoints, tower middleware, typed extractors.
+- Services (logical, not necessarily microservices):
+    - Identity/Auth service (local + optional ATProto OAuth integration)
+    - Content service (notes/cards/decks/sources)
+    - Study service (queue generation + grading + scheduling)
+    - Social service (follows, feeds, comments, notifications)
+    - Search service (indexing + query)
+    - Moderation service (reports, takedowns, rules)
+
+## Storage
+
+- Postgres: canonical app DB (users, private study state, cache of published records).
+- Object storage: images/audio, extracted article snapshots (if you store them).
+- Search index: separate system (Meilisearch/Typesense/ZincSearch-pick one later).
+
+## Eventing
+
+- Internal outbox pattern (DB table) for:
+    - reindex jobs, notification fanout, federation publish steps
+
+## Data Model
+
+See [Data Model Mapping](./data-model-mapping.md) for the mapping between public Lexicon records and internal database tables.
+
+### Core Entities
+
+- **Note**: markdown + structure + citations + links to sources.
+- **Card**: front/back (+ optional cloze, audio, image, code block).
+- **Deck**: ordered/clustered cards (+ metadata, tags).
+- **Lecture**: external URL + outline + timestamps + linked notes/cards.
+- **Article**: URL + extracted text (readability style heuristics) + highlights + linked notes/cards.
+- **Collection/Path**: curated bundle of decks + notes + sources.
diff --git a/docs/at-notes.md b/docs/at-notes.md
index c25d58d..0585a15 100644
--- a/docs/at-notes.md
+++ b/docs/at-notes.md
@@ -1,5 +1,7 @@
 # AT Protocol Research Notes
 
+Reference material for AT Protocol integration. For implementation details, see [todo.md](todo.md).
+
 ## OAuth 2.1 Specification
 
 AT Protocol uses a specific profile of OAuth 2.1 for client↔PDS authorization.
@@ -44,21 +46,22 @@ Format: `at://<did>/<collection>/<rkey>`
 
 Example: `at://did:plc:abc123/app.malfestio.deck/3k5abc123`
 
-## Firehose Consumption
+## Firehose / Jetstream
 
-For social features (trending, discovery, feeds):
+### Raw Firehose
 
-- **WebSocket Connection**: Subscribe to `com.atproto.sync.subscribeRepos` from a Relay
-- **CBOR Decoding**: Parse incoming events (or use Jetstream for JSON)
+- **WebSocket**: Subscribe to `com.atproto.sync.subscribeRepos` from a Relay
+- **CBOR Decoding**: Parse incoming events
 - **Cursor Management**: Track position for reconnection
 
-## AppView Pattern
+### Jetstream (Recommended)
 
-Index network-wide records to power discovery features:
+Bluesky's simplified JSON firehose:
 
-- Index `app.malfestio.*` records from firehose
-- Implement `getFeedSkeleton` for custom algorithmic feeds
-- Hydration service combines skeletons with full content from PDSes
+- JSON format (no CBOR decoding)
+- Reduced bandwidth (zstd compression)
+- Collection/repo filtering at source
+- Simpler reconnection with cursors
 
 ## Well-Known Endpoints
 
@@ -66,6 +69,46 @@ Index network-wide records to power discovery features:
 - `/.well-known/oauth-protected-resource` — PDS OAuth metadata
 - `/.well-known/oauth-authorization-server` — Auth server metadata
 
+## Labelers
+
+**Architecture:**
+
+1. Labels = metadata (source DID + subject AT-URI + value string)
+2. User Subscription = users subscribe to labelers; clients include in API requests
+3. Label Interpretation = per-user config to hide, warn, or ignore content
+
+**Structure:**
+
+```json
+{
+  "src": "did:plc:labeler",
+  "uri": "at://did:user/app.bsky.feed.post/123",
+  "val": "spam",
+  "cts": "2026-01-01T00:00:00Z"
+}
+```
+
+## Feeds
+
+**Core Flow**:
+
+1. User requests feed via at-uri of declared feed
+2. PDS resolves at-uri → Feed Generator's DID doc
+3. PDS sends `getFeedSkeleton` to service endpoint (authenticated by user's JWT)
+4. Feed Generator returns skeleton (list of post URIs + cursor)
+5. PDS hydrates skeleton with full content (via AppView)
+6. Hydrated feed returned to user
+
+## AppView
+
+**Responsibilities**:
+
+1. Record Processing & Indexing - consume firehose, build indices for likes, threads, follows
+2. Moderation Enforcement - apply labels from subscribed labelers
+3. Query Interface - expose XRPC API (proxied through PDS)
+4. Media CDN - fetch/cache blobs from upstream PDSes, generate thumbnails
+5. Search & Discovery - full-text search, type-ahead, content ranking
+
 ## Patterns from Real AT Protocol Apps
 
 ### plyr.fm (Music)
@@ -73,13 +116,11 @@ Index network-wide records to power discovery features:
 - OAuth 2.1 via `@atproto/oauth-client` library
 - Records synced to PDS: tracks, likes, playlists
 - Separate moderation service (Rust labeler)
-- Data ownership: "tracks, likes, playlists synced to your PDS as ATProto records"
 
 ### leaflet.pub (Writing)
 
 - React/Next.js frontend with Supabase + Replicache for sync
 - Bluesky integration via dedicated `lexicons/` and `appview/` directories
-- Publications posted to Bluesky
 
 ### wisp.place (Static Sites)
 
@@ -101,3 +142,6 @@ Index network-wide records to power discovery features:
 - [Repository & XRPC](https://atproto.com/specs/xrpc)
 - [Feed Generator Starter Kit](https://github.com/bluesky-social/feed-generator)
 - [atproto TypeScript SDK](https://github.com/bluesky-social/atproto)
+- [Ozone Moderation Service](https://github.com/bluesky-social/ozone)
+- [Jetstream Firehose](https://docs.bsky.app/blog/jetstream)
+- [Labels and Moderation Guide](https://docs.bsky.app/docs/advanced-guides/moderation)
diff --git a/docs/personas.md b/docs/personas.md
new file mode 100644
index 0000000..c5a5550
--- /dev/null
+++ b/docs/personas.md
@@ -0,0 +1,14 @@
+# Personas & Principles
+
+## Personas
+
+- **Learner**: studies daily; imports content; wants fast "review queue".
+- **Creator**: makes decks/notes; publishes updates; wants feedback + forks.
+- **Curator/Teacher**: bundles content into learning paths; annotates lectures/articles.
+- **Moderator/Community admin**: handles reports, takedowns, spam.
+
+## Design Principles
+
+- Local-first study experience; offline study must not feel "second-class".
+- Shareable artifacts are portable: Lexicon-defined schemas + stable IDs.
+- Privacy by design: progress + recall history are private unless explicitly shared.
diff --git a/docs/todo.md b/docs/todo.md
index 165bc08..5b792bb 100644
--- a/docs/todo.md
+++ b/docs/todo.md
@@ -21,7 +21,7 @@
 
 - Lexicon defines record types + XRPC endpoints; JSON-schema-like constraints.
 - Use "optional fields" heavily; avoid enums that will calcify the product too early.
-- Versioning: add fields, don’t rename; never rely on being able to rewrite history.
+- Versioning: add fields, don't rename; never rely on being able to rewrite history.
 
 ### Schema boundaries (important)
 
@@ -45,64 +45,255 @@
 - **(Done) Milestone D**: Identity + Permissions + Publishing Model.
     - Auth MVP, Permission model (Private/Public/SharedWith), and basic Publishing flow implemented.
     - Backend API and Frontend Editor updated with tests covering permissions and publishing.
+- **(Done) Milestone E**: Internal component library/UI Foundation + Animations.
 - **(Done) Milestone F**: OAuth + PDS Record Publishing.
     - OAuth 2.1 client flow (PKCE, DPoP, handle/DID resolution, token refresh).
     - PDS client for `putRecord`, `deleteRecord`, `uploadBlob`.
     - TID generation and AT-URI builder in core crate.
-- **(Done) Milestone E**: Internal component library/UI Foundation + Animations.
-- **(Done) Milestone F**: Content Authoring (Notes + Cards + Deck Builder).
-- **(Done) Milestone G**: Study Engine (SRS) + Daily Review UX.
+- **(Done) Milestone G**: Content Authoring (Notes + Cards + Deck Builder).
+- **(Done) Milestone H**: Study Engine (SRS) + Daily Review UX.
     - SM-2 spaced repetition scheduler.
-- **(Done) Milestone H**: Social Layer v1: Follow graph, Feeds (Follows/Trending), Forking workflow, and Threaded comments.
-- **(Done) Milestone I**: Search + Discovery + Taxonomy.
+- **(Done) Milestone I**: Social Layer v1: Follow graph, Feeds (Follows/Trending), Forking workflow, and Threaded comments.
     - Full-text search with pg_trgm/unaccent, visibility filtering, and unified search index.
     - Tag taxonomy and Discovery page with top tags.
 
-### Milestone J - Moderation + Abuse Resistance
+### Milestone J - Static Content & Landing Page
 
 #### Deliverables
 
-- Look into [Ozone](https://github.com/bluesky-social/ozone)
-- Reporting pipeline + review queue
-- Rate limits + spam heuristics
-- Takedown/visibility states (shadowed, removed, quarantined)
-- Audit logging for moderation actions
+**Marketing/Static Site:**
+
+- [ ] Landing page with hero section, feature highlights, social proof
+    - Hero should have a graph paper/grid background
+    - Floating flash cards, notes
+- [ ] "How it works" section with study flow visualization
+- [ ] About page with team/mission
+
+**App Vision Content:**
+
+- [ ] Onboarding flow with persona selection (Learner/Creator/Curator)
+- [ ] Empty states with helpful prompts for new users
+- [ ] Tutorial/walkthrough for first deck creation
+- [ ] Help center or FAQ section -> Should mention that the app is still in development and subject to change.
+
+**SEO & Meta:**
+
+- [ ] Open Graph / Twitter Card meta tags
+    - [ ] Scripted with HTML2Canvas/PNG generation
+        - Graph paper background
+        - Floating flash cards, notes
+- [ ] Sitemap.xml generation
+- [ ] robots.txt configuration
 
 #### Acceptance
 
-- You can safely operate an open publishing surface.
+- New visitors understand the value proposition within 10 seconds.
+- Onboarding flow guides users to create their first deck.
+
+### Milestone K - AppView Indexing
+
+#### Deliverables
+
+**Firehose Enhancement:**
+
+- [ ] Upgrade firehose consumer to store full record content (not just metadata)
+- [ ] Add `indexed_decks`, `indexed_cards`, `indexed_notes` tables for remote content
+- [ ] Track latest processed revision per repo; handle deletions
+
+**Search & Discovery:**
+
+- [ ] Implement search over indexed remote records (extend `search_items` view)
+- [ ] User profile aggregation: follower counts, deck counts from federated sources
+
+**Export & Interop:**
+
+- [ ] Export local records as valid Lexicon JSON (`/api/export/:collection`)
+- [ ] Read-only "federated library" view showing remote decks
+
+#### Acceptance
+
+- A deck published from Malfestio can be discovered via another AT Protocol client.
+- Remote decks from followed users appear in search results.
+
+### Milestone L - ATProto Integration Pass
+
+#### Deliverables
+
+**Identity & Auth:**
+
+- [ ] OAuth login directly to user's PDS (vs. local-only auth)
+- [ ] Handle resolution via DNS TXT or `/.well-known/atproto-did`
+- [ ] DPoP token binding for secure API calls
+
+**Sync & Conflict Resolution:**
+
+- [ ] Bi-directional sync: local drafts → PDS records, PDS records → local cache
+- [ ] Conflict resolution strategy for concurrent edits (last-write-wins or merge UI)
+- [ ] Offline queue for pending publishes
+
+**Deep Linking:**
+
+- [ ] AT-URI deep linking from external clients
+- [ ] Handle `at://` URL scheme in app
+
+#### Acceptance
+
+- User can log in with their existing Bluesky/PDS identity.
+- Local drafts sync correctly after reconnecting.
+
+#### Implementation Details
+
+**Considerations:**
+
+- Scalability: substantial compute; caching, DB optimization, distributed processing
+- Lexicon Validation: validate schemas, ignore invalid records gracefully
+- Account State: track latest processed revision per repo; handle deletions
+- Bluesky's AppView uses PostgreSQL or ScyllaDB + image proxy + AppView core
+
+**Identity:**
+
+- Use `did:web` for simplicity, `did:plc` for long-term stability
+- ATProto OAuth is the forward path
 
-### Milestone K - Federation / ATProto Integration Pass
+### Milestone M - Custom Feed Generator
 
 #### Deliverables
 
-- Phase 1 (minimum):
-    - export Lexicon records
-    - ingest remote records into a read-only "federated library"
-- Phase 2:
-    - OAuth login to PDS + publish records directly (client or server mediated)
-    - reconcile local drafts with remote published state
+**Infrastructure:**
+
+- [ ] Feed Generator service with `did:web` identity
+- [ ] Publish `app.bsky.feed.generator` declaration record to creator's repo
+- [ ] DID document with service endpoint for feed requests
+
+**Endpoints:**
+
+- [ ] `app.bsky.feed.getFeedSkeleton` - Return post URIs + cursor for pagination
+- [ ] `app.bsky.feed.describeFeedGenerator` - Feed metadata (DID, name, description)
+- [ ] JWT authentication for user-personalized feeds
+
+**Algorithms:**
+
+- [ ] "Trending Decks" - Top decks by fork/like count in last 7 days
+- [ ] "New from Following" - Latest decks from followed creators
+- [ ] "Study Streak Leaders" - Decks with highest completion rates (anonymized)
+
+**Indexing:**
+
+- [ ] Subscribe to `com.atproto.sync.subscribeRepos` (or Jetstream) for `app.malfestio.*` records
+- [ ] Index posts with compound cursor (timestamp::CID) for deterministic pagination
+- [ ] Garbage collect indexed data older than 48 hours (except pinned content)
 
 #### Acceptance
 
-- A published artifact is portable beyond your app.
+- Custom feed appears in Bluesky and other AT Protocol clients.
+- Feed surfaces relevant learning content based on engagement signals.
+- Pagination works correctly across feed refreshes.
+
+#### Implementation Details
+
+**Core Flow:** see [AT Notes](./at-notes.md#feeds)
+
+**Skeleton Response Format:**
 
-#### Notes
+```json
+{
+  "feed": [
+    {"post": "at://did:example/app.bsky.feed.post/1"},
+    {"post": "at://did:example/app.bsky.feed.post/2"}
+  ],
+  "cursor": "1683654690921::bafyrei..."
+}
+```
 
-- ATProto OAuth is the forward path; plan on it.
-- XRPC endpoint patterns and legacy session behavior exist, but treat them as transitional.
+**Skeleton Metadata Types:**
 
-### Milestone L - Reliability, Observability, Launch
+```typescript
+type SkeletonItem = {
+  post: string        // post URI
+  reason?: Reason     // optional context (e.g., repost)
+}
+type ReasonRepost = {
+  $type: 'app.bsky.feed.defs#skeletonReasonRepost'
+  repost: string      // repost URI
+}
+```
+
+**Considerations:**
+
+- Validate user JWTs if feed depends on user state (follows, likes)
+- Use compound cursor (timestamp::CID) for deterministic pagination
+- Most feeds can garbage collect data older than 48 hours
+- Reference: [Feed Generator Starter Kit](https://github.com/bluesky-social/feed-generator)
+
+### Milestone N - Reliability, Observability, Launch
 
 #### Deliverables
 
-- Metrics + tracing + structured logs
-- Backups + restore drills
-- Load test targets (study session + feed + search)
-- Beta program + feedback loop + roadmap iteration
+**Observability:**
+
+- [ ] Structured logging with correlation IDs
+- [ ] Metrics collection (Prometheus/OpenTelemetry)
+- [ ] Distributed tracing for request flows
+- [ ] Error tracking (Sentry or similar)
+
+**Reliability:**
+
+- [ ] Database backups + restore drills
+- [ ] Health check endpoints (`/health`, `/ready`)
+- [ ] Graceful shutdown handling
+- [ ] Circuit breakers for external dependencies
+
+**Load Testing:**
+
+- [ ] Study session throughput targets
+- [ ] Feed generation latency benchmarks
+- [ ] Search query performance under load
+
+**Launch Prep:**
+
+- [ ] Beta program signup flow
+- [ ] Feedback collection mechanism
+- [ ] Feature flags for gradual rollout
+
+#### Acceptance
+
+- System handles 10x expected load without degradation.
+- Mean time to recovery < 5 minutes for common failures.
+
+### Milestone O - Moderation + Abuse Resistance
+
+#### Deliverables
+
+**Labeler Infrastructure:**
+
+- [ ] Dedicated Bluesky account for labeler service
+- [ ] Publish `app.bsky.labeler.service` record to make discoverable
+- [ ] Self-host Ozone backend + UI (Docker setup in `HOSTING.md`)
+- [ ] Configure report types via `goat` CLI
+
+**Endpoints:**
+
+- [ ] `com.atproto.label.subscribeLabels` - real-time label stream
+- [ ] `com.atproto.label.queryLabels` - query published labels
+- [ ] `com.atproto.report.createReport` - accept user reports
+
+**Moderation Features:**
+
+- [ ] Reporting pipeline + review queue UI
+- [ ] Rate limits + spam heuristics
+- [ ] Takedown/visibility states (shadowed, removed, quarantined)
+- [ ] Audit logging for moderation actions
+
+#### Acceptance
+
+- You can safely operate an open publishing surface.
+- Users can subscribe to your labeler and see moderation applied.
+
+**Reference:** [Ozone Moderation Service](https://github.com/bluesky-social/ozone)
 
-## Open Questions (Parked Decisions)
+## Open Question/Parked Decisions
 
-- Local-first mechanics: full offline authoring + later publish, or online-only creation?
-- Federation depth: read-only ingest first, or publish-to-PDS in the first public beta?
-- Content extraction: store extracted article snapshots (legal/ops implications), or store only metadata + highlights?
+- Full offline authoring + later publish
+- Federation depth: publish-to-PDS in the first public beta
+- Content extraction: store extracted article snapshots locally (browser)
+    - Persist only metadata + highlights