Skip to content

Latest commit

 

History

History
444 lines (336 loc) · 19.6 KB

File metadata and controls

444 lines (336 loc) · 19.6 KB

Strata — Observations & Future Roadmap

Last Updated: April 15, 2026


UX/UI Observations (Near-term)

1. Typography & Readability

Status: Flagged for batch UX pass
Issue: Serif font looks great on landing page but becomes difficult to read at smaller sizes in admin screens.
Action: Audit all typography hierarchy; consider switching to sans-serif for dense admin UI while keeping serif for marketing/public-facing pages.
Priority: Medium (impacts daily admin workflow)


Feature Requests (Pipeline Enhancements)

2. Partial Processing / Incremental Segmentation

Status: Phase 2 feature
Request: Process books in chunks (e.g., chapters 1-2 now, chapters 3-4 later) rather than all-or-nothing.
Context: Users may want to process a few chapters, review quality, then do more.
Implementation:

  • Add startChapter and endChapter params to pipeline job
  • Modify segmentation logic to extract only specified range
  • Auto-calculate maxChars based on chapter range

Priority: High
Complexity: Medium (requires DB schema extension, pipeline refactor)


3. Chapter/Section Selection UI

Status: Phase 2 feature
Request: Instead of dropdown with fixed options, allow users to specify exact chapter ranges (or entire book).
Current: Limited to predefined maxChars options
Desired: "Process chapters 5-12" or "Process entire book" or "Preview first 3 chapters"

Implementation:

  • Fetch chapter metadata from Gutenberg metadata API or parse from book text
  • Build UI to show chapter list + checkboxes or range selector
  • Pass chapter indices to pipeline

Priority: High
Complexity: Medium–High (requires book structure analysis)


4. PKM Integrations (Notion, Obsidian, Logseq, Apple Notes)

Status: Phase 2+ feature
Request: Export user highlights, notes, and commentary to PKM systems.
Use Case: Readers want their study notes to flow into their personal knowledge base.
Possible Integrations:

  • Notion: API-based export (highlights as database, commentary as linked pages)
  • Obsidian: Markdown export with wikilinks to episodes
  • Logseq: Outline format with backlinks
  • Apple Notes: iCloud sync
  • Generic: Markdown dump that user can import anywhere

Implementation Approach:

  1. Add /api/export/notion, /api/export/obsidian, etc.
  2. Format user's notes + highlights + AI commentary as destination format
  3. Handle auth (Notion token, Obsidian vault path, etc.)
  4. Provide one-click "Export to Notion" in UI

Priority: Medium (nice-to-have, differentiator)
Complexity: Medium (requires learning each platform's API)
Value: Increases stickiness; turns Strata into part of user's workflow


ElevenLabs Audiobook Generator: Strategic Decision

Current State

  • Strata uses ElevenLabs TTS (single voice, eleven_multilingual_v2)
  • Narrates at ~30–50 words per minute depending on voice
  • Quality is good but sounds synthetic

ElevenLabs Audiobook Generator

  • Professional audiobook-grade narration (likely human-quality)
  • Character voices (different speakers per character in dialogue)
  • Pacing & inflection tuned for literary content
  • Premium pricing (presumably higher cost than TTS API)

Two Strategies

Strategy A: Replace TTS with ElevenLabs Audiobook Generator

Pros:

  • Dramatically better audio quality (human-like narration)
  • Industry-standard for publishing
  • Differentiates from generic TTS
  • Makes audiobooks sellable on other platforms

Cons:

  • Higher API costs
  • May not support per-episode generation (might require full book at once)
  • Less control over voice selection
  • Vendor lock-in to ElevenLabs

Recommendation: Good if you want premium distribution-quality audiobooks.


Strategy B: Use Both (Hybrid Approach)

Strata for Free Users: Fast TTS narration (current)
Strata Pro / ElevenLabs Audiobook: Premium narration via ElevenLabs Generator
Then Distribute: Export both versions to ElevenReader marketplace

Pros:

  • Scalable (free users get quick preview, paid users get pro narration)
  • Monetization layer (charge for premium narration)
  • Flexible (user chooses quality vs. speed tradeoff)

Cons:

  • More complex pipeline
  • Higher support burden

Recommendation: Best if you want to offer tiers.


Strategy C: Full Distribution Play

  1. Process books in Strata (commentary, segmentation, themes)
  2. Narrate with ElevenLabs Audiobook Generator (professional quality)
  3. Publish to:
    • ElevenReader marketplace (60% revenue)
    • Apple Books, Audible, Scribd (through aggregators like Draft2Digital)
    • Your own Strata app (subscription model)
  4. Own the metadata & relationship — you control the experience

Pros:

  • Maximize revenue streams (SaaS + marketplace + wholesale)
  • Use ElevenLabs as a service tier, not the main product
  • Leverage your annotation work across multiple channels
  • Could become a full publishing operation

Cons:

  • More complex (requires aggregator integrations)
  • Cash flow timing (payments take 90+ days via aggregators)

Recommendation: Long-term play for scale.


My Take

Replace the current TTS with ElevenLabs Audiobook Generator as Phase 2.

Here's why:

  1. Your current TTS is fine for MVP but won't compete with real audiobooks
  2. ElevenLabs Audiobook Generator is better because that's what it's designed for
  3. The better the audio, the more you can charge/distribute
  4. You're not building an audio engine; you're building a learning platform

Don't try to be better at narration than ElevenLabs. Use their engine. Your edge is:

  • Episode structure + commentary
  • Highlights + annotation
  • Learning journey (not just listening)
  • Community features (chat, notes)

Suggested Roadmap:

  1. Now: Ship MVP with current TTS (it works)
  2. Phase 1: Chapter selection, partial processing
  3. Phase 2: Switch to ElevenLabs Audiobook Generator for better quality
  4. Phase 2: ElevenReader marketplace export
  5. Phase 2–3: PKM integrations (Notion export)
  6. Phase 3+: Direct distribution (Apple Books, Audible via aggregators)

The key insight: You're not in the narration business. You're in the literary curation + learning design business. ElevenLabs should handle audio.


Batch Work Items (Next Sprint Planning)

Context

ElevenLabs recently launched ElevenReader — an audiobook marketplace where creators can:

  • Upload audiobooks
  • List them on ElevenReader marketplace
  • Earn 60% of revenue
  • Leverage ElevenLabs TTS infrastructure

Question: How does Strata differentiate, and can we integrate?


Strata's USP vs. ElevenReader

Aspect ElevenReader Strata
Core Function Audiobook distribution & hosting Content segmentation + guided literary experience
Content Processing Upload pre-existing audio Cut books into thematic episodes w/ commentary
Context Layers Just audio Episodes + notes + highlights + AI commentary + quote cards + chat
User Experience Listen & navigate Structured learning journey with interleaved study aids
Monetization 60% revenue share Direct subscription to readers (+ marketplace option)
Customization Publisher decides structure Strata algorithmically segments + AI annotates
Target Audience General audiobook listeners Deep readers, learners, Great Books fans, Catholics
Workflow Upload final product Source text → pipeline → annotated episodes → publish

Strata's Unique Strengths

  1. Structured Learning — Episodes aren't arbitrary; each has commentary, themes, character arcs
  2. AI-Powered Annotation — Chat, quote extraction, thematic abridgments aren't available elsewhere
  3. Niche Focus — Religious, classics, Great Books positioning (not generic audiobooks)
  4. Direct Relationship — You own the reader experience; ElevenReader is just distribution
  5. Data Leverage — User highlights, notes, chat give you insight into what resonates
  6. Flexible Monetization — Subscription app + selling individual titles to libraries + ElevenReader marketplace

Integration Opportunity: Strata → ElevenReader Export

Scenario

  1. User processes a book in Strata (e.g., Brothers Karamazov)
  2. Strata generates annotated episodes with narration
  3. Export to ElevenReader: Package as EPUBs/PDFs with audio links + metadata
  4. List on ElevenReader marketplace
  5. Earn 60% from listeners who discover via ElevenReader

Benefits

  • Reach: ElevenReader brings discovery (you handle production)
  • Revenue: 60% is pure upside on top of Strata subscription model
  • No Conflict: You're not competing with ElevenLabs; you're using their distribution
  • Differentiation: Your commentary + curation makes your books better than generic TTS audiobooks

Implementation Roadmap

  1. Phase 1: Ensure Strata pipeline exports clean EPUB + audio files
  2. Phase 2: Build ElevenReader export connector (package episode metadata + links)
  3. Phase 3: Wire ElevenReader earnings dashboard into Strata admin
  4. Phase 4: Marketing: "Books on Strata + ElevenReader marketplace"

Recommended Strategy

Keep building Strata as the primary product. Your core USP is the guided learning experience, not just audiobooks. ElevenReader becomes a distribution channel, not a competitor.


Distribution Channels & Platform Strategy

Current / Planned Channels

Channel Content Type Model Monetization Primary Use
Strata App Audio + text + learning layer Freemium subscription Direct (free trial → $X/mo) Core product
ElevenReader Audio + metadata Free listing 60% revenue split Discovery + audiobook market
YouTube Audio + transcript overlay Free content Ad-supported or sponsorship Discovery + marketing
Podcast Curated audio commentary Free feed Ad-supported or sponsorship Discovery + audio format
Substack Weekly written commentary Free newsletter Email list building Top-of-funnel marketing
Beehiiv Weekly written commentary Free newsletter Email list building + potential affiliate Top-of-funnel marketing
PKM Exports Written commentary + notes API/integration Value-add to Strata Pro Workflow integration

Substack Analysis: Free Newsletter vs. Paid

Substack could work, but only if positioned correctly:

❌ Substack as Competing Paid Subscription

  • Another paywall (fractures audience)
  • Competes with Strata Pro (confusion)
  • Audio/video are secondary on Substack (not ideal)
  • Don't do this — you'd be splitting revenue and attention

✅ Substack as Free Weekly Email

  • Free weekly "commentary digest" (e.g., top 3 quotes + themes from the book)
  • Drives discovery to Strata app
  • Builds email list (owned asset)
  • Better than podcast/YouTube for written content + email-native audience
  • Low effort (repurpose commentary already generated)

Recommended Distribution Stack

Priority 1: Strata App (core product, direct revenue)
Priority 2: Podcast (free, audio-first, easy distribution)
Priority 2: YouTube (free, algorithmic discovery, ads/sponsorship)
Priority 3: Email Newsletter (free Substack digest, drives to Strata)
Priority 3: ElevenReader (60% splits, passive revenue)
Priority 4: PKM Exports (Notion/Obsidian, value-add)

Why Not Pay-Gated Substack

  1. Audience fragmentation: You'd have Strata subscribers + Substack subscribers (confusing)
  2. Substack doesn't do audio well (their platform is text-first, not audio-first)
  3. Substack takes 10% of revenue (vs. 0% on your own app)
  4. Email is a weak substitution for app (notification, features, learning structure)

Why Free Substack Email Could Work

  1. Owned email list (Substack makes it easy, but it's your data)
  2. Low-effort repurposing (commentary already exists, just digest it)
  3. Top-of-funnel marketing (free weekly → "Want full guided experience? Try Strata Pro")
  4. Complements podcast/YouTube (different format, email-native audience)
  5. Drives traffic to Strata (not competing for subscription revenue)

Dual Newsletter Strategy: Substack + Beehiiv

Why run both?

  • Different audiences: Some readers prefer Substack ecosystem, others prefer Beehiiv
  • Same content: Single source (your commentary digest) → auto-publish to both platforms
  • Redundancy: If one platform has issues, you still have reach on the other
  • SEO/discovery: Beehiiv has built-in sponsor marketplace (potential affiliate revenue from book recommendations)
  • Low marginal cost: Once you digest the commentary, distributing to both is minimal extra effort

Implementation:

  • Write once (weekly digest)
  • Publish to Substack and Beehiiv simultaneously
  • Include CTA in both: "For full guided experience, listen on Strata"
  • Track clicks/conversions from each to see which drives more Strata signups

Note: Both should remain FREE (no paywall comparison between platforms).

Suggested Approach

Later (Phase 3+): Launch free weekly Substack digest that mirrors your podcast/YouTube to capture email-native audience. Subject line: "Top 3 insights from [Book Chapter]" → links to full guided experience on Strata.

Now: Focus on podcast + YouTube + Strata app.


The Key Insight

If free channels (ElevenReader, YouTube, podcast) have poor audio quality, listeners won't even consider Strata. Audio quality is your acquisition hook, not your paywall.

Conversely, if they experience premium narration for free, they want to share it — organic growth.

Corrected Freemium Model

Channel Audio Quality Added Value Model
ElevenReader Marketplace Premium (Audiobook Generator) None Free / 60% splits
YouTube Premium (Audiobook Generator) Transcript overlay Free / Ad-supported
Podcast Premium (Audiobook Generator) Curated episode selection Free / Ad-supported
Strata App (Free Trial) Premium (same) Audio + basic features 2-week free trial
Strata App (Pro Subscription) Premium (same) + Guided commentary + highlights + chat + notes + structure Paid

Why This Works

The paywall is NOT audio quality. It's the learning layer.

Free users get: "Wow, this audiobook is beautifully narrated."
Paid users get: "Wow, AND I understand it deeply with guided commentary and community."

Audio Strategy

  • Everywhere (free + paid): Premium narration via ElevenLabs Audiobook Generator (or professional narrator)
  • NOT a paywall lever
  • Differentiation: Your commentary + structure + chat + annotations = the paid experience

Freemium Conversion Funnel

  1. Discover → ElevenReader / YouTube / podcast (premium audio, no paywall)
  2. Download Strata → Try free trial (same premium audio)
  3. Hit study feature → "Want guided commentary?" (paywall)
  4. Subscribe → Unlock full learning experience (same audio + structure)

Benefits of This Approach

  • Word-of-mouth: "This audiobook is incredibly well narrated" (true, no TTS artifacts)
  • No friction: Audio is the same paid/free
  • Conversion happens on learning value (commentary, structure, community), not on audio gatekeeping
  • Scalable: You use premium narration everywhere; differentiation is curation + learning design
  • Competitive: Your position is "Great audio + Great learning", not "OK audio locked behind paywall"

Recommendation: Proceed with Strata, Plan ElevenReader as Phase 2 Distribution

Why

  1. ElevenReader is a commodity marketplace — success depends on content quality & marketing
  2. Strata is a unique platform — no competitor offers structured learning + AI annotation for classics
  3. Strata subscription captures direct value — ElevenReader is gravy on top
  4. You own the brand & data — readers getting smarter is your asset; ElevenReader just helps distribute

Go-Forward Plan

Now (MVP): Perfect the Strata experience on Brothers Karamazov
Next (Phase 1): Add chapter selection, partial processing
📋 Later (Phase 2): ElevenReader export + marketplace listing
📋 Eventually: Native Strata publishing (books → readers directly without ElevenReader)


5. Intro/Discovery Messaging for Exported Audiobooks

Status: Phase 2+ feature
Request: Add a preamble to each chapter (audio or text) directing listeners to deeper experience.
Example: "You're listening to Chapter 5 of Brothers Karamazov on [Platform]. For guided commentary, notes, and deep discussion, listen with Strata.com or subscribe to our podcast/YouTube channel."

Implementation:

  • Add optional introMessage field to episodes table
  • Template: "Listen with Strata at [link] | Subscribe to podcast: [link] | Watch on YouTube: [link]"
  • When exporting to ElevenLabs Audiobook Generator, prepend intro to episode narration
  • Also export in text formats (EPUB, Markdown)

Priority: Medium (good for conversion but not blocking)
Complexity: Low (just text field + template logic)
Value: Cross-promotes your other channels; drives conversion from audiobooks → Strata subscription


6. Decouple Text Generation from Audio Generation (Pipeline Refactor)

Status: Architecture decision for Phase 2
Current: Pipeline does everything end-to-end (segment → commentary → abridgments → audio)
Proposed: Two-step workflow

Step 1: Text-Only Export

  • Generate commentary, abridgments, highlights, quote cards
  • Export as: JSON metadata, Markdown, EPUB, plain text
  • Output: Ready for external narration services (ElevenLabs Generator, human narrators, etc.)
  • Use cases:
    • Distribute commentary separately from audio
    • Archive text layers without audio
    • Pass to ElevenLabs Audiobook Generator (bring-your-own-narration)
    • Publish as ebooks independently

Step 2: Audio Generation (Optional Later)

  • Still in pipeline but now decoupled
  • Can apply to any text export
  • User chooses: use ElevenLabs TTS, swap in Audiobook Generator, or skip entirely

Admin Portal Flow:

  1. Upload/fetch book
  2. Trigger: "Generate Text Only" → commentary, abridgments, notes
  3. Review & export as JSON/EPUB/Markdown
  4. (Optional later) Trigger: "Generate Audio" → choose narrator (TTS / Audiobook Generator / other)

Benefits:

  • You own the text layers (commentary, structure, curation)
  • Audio is now a separately monetizable/distributable asset
  • Can use different narrators for different channels
  • Supports multiple export formats for PKM/publishing
  • Faster initial processing (no waiting for audio)

Database Changes:

  • Add textProcessingStep status to pipelineJobs
  • Add audioProcessingStep status (separate)
  • Export formats stored in S3 (JSON, EPUB, Markdown alongside MOBI/PDF)

Implementation Complexity: Medium–High (requires pipeline refactor)
Priority: High (unlocks ElevenLabs Audiobook Generator integration + PKM exports)


Batch Work Items (Next Sprint Planning)

  • Typography audit: serif → sans-serif for admin, keep serif for marketing
  • Add startChapter/endChapter to pipeline jobs schema (migration)
  • Build chapter selection UI for pipeline trigger
  • Research ElevenReader API/export requirements
  • Investigate Gutenberg chapter metadata API
  • Spec out partial processing logic
  • Design text-only export workflow + formats (JSON, EPUB, Markdown)
  • Refactor pipeline: decouple text generation from audio generation
  • Add intro/discovery messaging template to episodes