Lemons Business Plans #001 - Language Learning on NOSTR

The future of language tutoring is not a better platform. It is no platform at all. It is an open protocol, and the best tools on top of it.
Lemons Business Plans #001 - Language Learning on NOSTR

This Business Plan is open source - feel free to use it - WTFPL


This business plan describes a protocol-first approach to language tutoring marketplaces. All specifications (event kinds, NIPs) are proposed and would require community discussion and standardization. Financial projections are illustrative. The Nostr protocol and Lightning Network are decentralized technologies and Lumen as described would be one of many potential participants in their ecosystems.

Business Plan: Lumen — An Open-Protocol Language Tutoring Marketplace Built on Nostr


1. Executive Summary

Lumen is an open, permissionless marketplace for language tutoring built on the Nostr protocol (Notes and Other Stuff Transmitted by Relays). Unlike centralized platforms like Preply that extract 18–33% of every tutor’s earnings, Lumen establishes a protocol-first architecture where tutors and students transact directly, reputation is cryptographically portable, and no single entity controls access, pricing, or relationships.

The business does not own the marketplace — it serves it. Revenue is generated by building premium tools on top of the open protocol: a best-in-class client application, a purpose-built tutoring relay with advanced discovery, AI-powered teaching co-pilots, corporate management dashboards, and optional dispute resolution services.

For tutors, the value proposition is radical: keep ~99%+ of what you earn instead of 67–82%. For students, prices fall while quality rises as competition becomes truly open. For Lumen, the business model shifts from rent extraction to genuine value creation — and captures a fraction of a much larger, more defensible ecosystem.


2. Why Nostr? Why Open Protocols?

2.1 The Problem with Centralized Tutoring Marketplaces

Centralized Model (Preply/itaki) Problem
Platform owns tutor-student relationship Tutors cannot export reviews, student history, or reputation
18–33% commission on all lessons A tutor earning $40K/year gives $7K–$13K to the platform
Deplatforming risk Tutors removed for policy violations, real or alleged, lose entire income stream overnight
Opaque algorithms Search ranking is a black box; tutors cannot audit why they appear or disappear
Data silos Student learning data, reviews, and progress are locked into one platform
Single point of failure If the company shuts down, the marketplace disappears

2.2 The Nostr Advantage

Nostr is a decentralized protocol for social communication where users publish cryptographically signed messages (“events”) to relays — servers that anyone can run. Key properties that make it ideal for a marketplace:

  • Self-sovereign identity: Every user controls a public/private keypair. No email, no phone number, no platform account required. Identity is portable.

  • No central server: Anyone runs relays. The marketplace exists wherever relays exist. No single company can shut it down.

  • Cryptographic reputation: Reviews, ratings, and transaction history are signed events. They cannot be forged, and they follow the user — not the platform.

  • Native payments: Lightning Network micropayments are already deeply integrated into the Nostr ecosystem via “zaps.” Payment is a protocol primitive, not a bolt-on.

  • Open clients: Anyone can build an app that reads from and writes to the same relays. Competition happens at the client layer, not through walled gardens.

  • Composable: The tutoring marketplace can interoperate with Nostr-based identity (NIP-05), calendar tools, messaging, and content feeds.

The core insight: A tutoring marketplace is fundamentally about (a) discovering a tutor, (b) booking a session, (c) conducting the session, (d) paying, and (e) reviewing. Every one of these is a discrete protocol action that can be expressed as Nostr events.


3. Protocol Architecture — The Open Marketplace

3.1 Nostr Event Kinds for Tutoring

The entire marketplace operates through a defined vocabulary of Nostr event kinds. These are open specifications — anyone can implement them.

Action Nostr Event Kind Content Proposed NIP
Tutor Profile Kind 30XYZ Language(s) taught, hourly rate (sats/fiat), bio, video intro URL, accent/dialect, availability, credentials Custom NIP
Calendar/Availability Kind 31XYZ JSON array of available time slots, timezone Custom NIP
Booking Request Kind 4 or 14 (encrypted DM) Proposed time, duration, language, student notes NIP-04/NIP-17
Booking Confirmation Kind 32XYZ Accepted time, Lightning invoice, session metadata Custom NIP
Payment Kind 9735 (Zap Receipt) Proof of payment for session NIP-57
Lesson Completed Kind 33XYZ Session metadata, duration, notes Custom NIP
Review/Rating Kind 1 or custom 34XYZ Rating (1–5), text review, rebooking intent Custom NIP (extends NIP-01)
Credential/Verification Kind 35XYZ TEFL certificate hash, degree credential, background check status Custom NIP
Dispute Kind 36XYZ Encrypted dispute details to optional arbiter Custom NIP
Tutor Availability Status Kind 30311 (Live Event) or custom “Available now” / “Offline” / “Accepting students” Extends NIP-53

3.2 How a Transaction Flows (Protocol View)

1. Tutor publishes Kind 30XYZ (profile + offerings) → relays
2. Student's client queries relays for "Spanish tutors under 20k sats/hour"
3. Student browses profiles, reads reviews (Kind 34XYZ events), selects tutor
4. Student sends encrypted DM (Kind 14) with booking request + preferred time
5. Tutor receives DM, sends back Kind 32XYZ confirmation with Lightning invoice
6. Student pays Lightning invoice instantly (milliseconds, <0.1% fee)
7. Video session happens (WebRTC, Jitsi, or Whereby — client handles this)
8. After session, student publishes Kind 34XYZ review (signed, immutable)
9. Tutor publishes Kind 33XYZ session record (creates on-chain teaching history)

No platform in the middle. No commission. No permission.

3.3 Identity & Trust

NIP-05 Verified Identities: tutor@lumen.tutoring — a human-readable identifier tied to the tutor’s public key, hosted on Lumen’s domain (or any domain the tutor chooses). This is the Nostr equivalent of a verified badge.

Web of Trust: Reviews are signed by the reviewer’s key. A student can verify that reviews come from real students (with transaction history on-chain) — no fake reviews possible without detectable Sybil behavior.

Credential Verification Service: Lumen runs an optional credential attestation relay. Tutors submit credentials; Lumen verifies and publishes a signed attestation event. The tutor can take this attestation anywhere.

Portable Reputation: A tutor who has accumulated 200 five-star reviews and 500 completed lessons on Lumen’s relay can move to any other client or relay and bring that entire history with them. The data lives on relays; the reputation lives with the tutor.


4. The Business Layer — How Lumen Makes Money

The protocol is free and open. Lumen the company builds value on top and charges for it. This is the “pick-and-shovel” model of an open marketplace.

4.1 Revenue Streams

Revenue Stream Description Pricing % of Revenue (Y3 Target)
Lumen Client (premium) Best-in-class mobile & web app for tutors and students Freemium: free basic client; $4.99/month Pro (AI co-pilot, advanced analytics, priority relay, no ads) 35%
Lumen Relay (hosted) High-performance, tutoring-optimized relay with search indexing, spam filtering, and guaranteed uptime Free read; $9.99/month for tutor write access (ensures relay quality); enterprise relay $199/month 25%
AI Tutor Co-Pilot Lesson planner, pronunciation feedback, vocabulary builder, progress reports, student homework generation Included in Pro; standalone $7.99/month 15%
Corporate Dashboard (B2B) Manage employee language training across the open protocol; bulk Lightning payments, progress dashboards, HR reporting $499–$2,999/month based on seats 15%
Dispute Resolution & Escrow Optional third-party mediation for disputes; Lightning-native escrow contracts (Hodl invoices) 2.5% of disputed transaction value (only when used) 5%
Verification Service Background checks, credential verification, identity attestation published to relays $49 one-time per tutor 3%
White-Label Client Universities and language schools run their own branded client on the shared open marketplace $1,500 setup + $499/month 2%

Total addressable revenue is fee-for-service, not rent-extraction. No one is forced to pay Lumen anything to participate in the marketplace. But Lumen builds products so good that users choose to pay.

4.2 Why This Beats the Commission Model

Centralized (Preply) Open Protocol (Lumen)
Tutor keeps 67–82% 99.4%+ (only Lightning routing fee ~0.01–0.1%)
Student hourly cost (for a $20/hour tutor) $20 $14–$16 (tutors can charge less and still earn more)
Reputation portability None — locked to platform Full portability across clients and relays
Deplatforming risk High Zero — data and relationships live on relays
Censorship resistance Low High — no central authority
Innovation speed Single company bottleneck Permissionless — anyone builds on the protocol
Trust model “Trust the platform” “Trust math” — cryptographic verification

5. Product & Client Architecture

5.1 The Lumen Client

The flagship product is a polished, user-friendly client application (web, iOS, Android) that serves as the primary interface to the open marketplace. It is one of potentially many clients — competition improves quality for everyone.

Free Tier:

  • Browse tutors, read reviews, book sessions

  • In-app video classroom (Jitsi Meet integration)

  • Basic Lightning wallet (Nostr Wallet Connect)

  • Publish reviews

  • Standard relay support

Pro Tier ($4.99/month):

  • AI-powered tutor matching

  • Integrated AI lesson co-pilot (plan suggestions during sessions)

  • Advanced analytics (learning progress, hours tracked, spending)

  • Priority relay access (faster search, better uptime)

  • Ad-free experience

  • Multi-language learning paths

  • Session recording with AI-generated notes

  • Export all data (full data portability — a protocol-native right)

5.2 The Lumen Relay

A purpose-built Nostr relay optimized for tutoring marketplace events:

  • Full-text search over tutor profiles, languages, and reviews (Elasticsearch-backed)

  • Spam filtering: Rate limiting, proof-of-work requirements for new keys, reputation-weighted event acceptance

  • Event indexing: Fast queries for “Spanish tutors in my timezone under 20k sats”

  • Guaranteed uptime SLA for paid tutors (99.9%)

  • GDPR-compliant deletion (tutors can request event deletion; relay honors it while events persist on other relays they’ve published to)

  • NIP-42 authentication for paid access

5.3 AI Co-Pilot (In-Client & Standalone)

  • Pre-session: Generates a personalized lesson plan based on student’s stated goals, past session notes, and vocabulary gaps

  • During session: Real-time pronunciation feedback (speech-to-text + phonetic analysis), vocabulary logging

  • Post-session: Auto-generated session summary, flashcards from new vocabulary, homework assignments

  • Long-term: Progress dashboards, CEFR level estimation, recommended next tutor (if current tutor isn’t available)

The AI co-pilot reads from the open protocol (session history, reviews, student notes) and writes back structured data. It is a client-side enhancement — no server-side data lock-in.

5.4 Corporate B2B Dashboard

Companies purchase Lightning-hosted lesson packages for employees. The dashboard provides:

  • Employee enrollment and seat management

  • Bulk Lightning payment automation

  • Learning progress dashboards across teams

  • Compliance reporting (for companies with L&D mandates)

  • Multi-tutor management (employees can use any tutor on the open marketplace)

  • SSO integration (SAML/OIDC)

The corporate client pays Lumen for the dashboard — the tutors still receive direct payments. Lumen does not intermediate the tutoring relationship.


6. Technical Architecture

6.1 Stack Overview

┌─────────────────────────────────────────────────┐
│                  CLIENT LAYER                     │
│  (Lumen App, Third-Party Apps, Corporate Client)  │
├─────────────────────────────────────────────────┤
│              SERVICE LAYER (optional)              │
│  AI Co-Pilot • Verification • Dispute • Search   │
├─────────────────────────────────────────────────┤
│                 RELAY LAYER                        │
│  (Lumen Relay, Public Relays, Personal Relays)    │
├─────────────────────────────────────────────────┤
│               PAYMENT LAYER                        │
│  (Lightning Network • Bitcoin L1 • Stablecoins)   │
├─────────────────────────────────────────────────┤
│               IDENTITY LAYER                       │
│  (Nostr Keys • NIP-05 • DID • LNURL-auth)         │
└─────────────────────────────────────────────────┘

6.2 Technology Choices

Component Technology Rationale
Client (Web) React/Next.js, TailwindCSS Modern, fast, SEO-friendly
Client (Mobile) React Native or Flutter Shared business logic with web; native performance
Nostr SDK NDK (Nostr Development Kit) or nostr-tools Mature, well-maintained libraries
Relay Strfry or Nostr-rs-relay (Rust) High-performance, low-resource relay implementations
Search Index Elasticsearch or Meilisearch (sidecar to relay) Full-text search over tutoring events
Lightning LND, Core Lightning, or Alby/LNBits Self-hosted or managed Lightning node
Nostr Wallet Connect NIP-47 Decoupled Lightning payments from client
Video Jitsi Meet (self-hosted or managed) Open-source, no vendor lock-in, WebRTC
AI/LLM Open-weight models (Llama, Mistral) or OpenAI API Lesson planning, pronunciation analysis
Authentication NIP-07 (browser extension) or NIP-46 (remote signer) Secure key management without exposing private key
Data Storage Relay events + optional client-side encrypted DB All marketplace data lives on relays

6.3 Key Protocol Design Decisions

Why Lightning, not fiat? The Nostr ecosystem’s payment primitive is the Lightning Network. This provides instant, near-zero-fee global settlement. However, most students and tutors are not Bitcoin-native. The Lumen client abstracts this:

  • Students pay via credit card → Lumen’s payment processor converts to sats → Lightning payment to tutor

  • Tutors receive sats → optional auto-conversion to local currency (via Strike, Breez, or similar Lightning-to-fiat services)

  • The protocol records everything in sats; the UI shows local currency

Over time, as users hold and reuse sats within the ecosystem, conversion friction disappears.

Why not a blockchain/smart contract approach? Nostr is deliberately not a blockchain. It has no consensus mechanism, no gas fees, no on-chain storage costs, and no scaling bottlenecks. Events are gossiped between relays — the architecture is more like email or RSS than Ethereum. This is the correct abstraction for a marketplace where most data (profiles, reviews, availability) does not require global consensus.


7. Go-To-Market Strategy

7.1 The Cold Start Problem — Solved Differently

Traditional marketplaces face a chicken-and-egg problem. Open protocols face a bootstrapping problem: who publishes events to an empty relay?

Strategy: Seed the marketplace with an initial cohort of high-quality tutors.

  1. Recruit 200 tutors from existing Nostr-adjacent communities: Bitcoiner language tutors, pleb.fi community, Stacker News users. These individuals are already comfortable with Lightning and self-custody.

  2. Onboard them with white-glove service: Lumen team helps set up Nostr keys (NIP-07), configures profiles, verifies credentials, and provides a guaranteed first 10 students (via paid acquisition).

  3. Subsidize early student demand: First 1,000 students get 3 free lessons (paid for by Lumen’s marketing budget). This is far cheaper than centralized platforms’ CAC because there’s no ongoing commission to subsidize — it’s a one-time acquisition cost.

  4. Language-specific launch: Start with English tutoring from the Philippines and Spanish tutoring from Mexico/Colombia. These are high-demand, high-supply language pairs with significant geographic arbitrage.

7.2 Channel Strategy

Channel Message Target
Nostr-native channels “Tutors: Keep 99% of what you earn. Own your reputation. No deplatforming.” Early tutor supply
Lightning/Bitcoin communities “Earn sats teaching your native language” Tutor supply, early student demand
Tutor communities (Reddit, Facebook groups) “Stop giving 33% to platforms. Teach on an open marketplace.” Tutor supply
Language learning communities (Reddit, Discord) “Learn from native speakers for 30% less. No platform markup.” Student demand
YouTube/TikTok Tutor testimonials: “I used to give Preply $800/month. Now I keep it all.” Mass-market awareness
B2B outreach “Corporate language training on open protocols — auditable, portable, no lock-in” Enterprise revenue

7.3 Brand Positioning

Centralized platforms stand for “trust us.”

Lumen stands for “trust the protocol.”

The brand is built on sovereignty, transparency, and economic fairness. The core narrative: “Your relationship with your tutor belongs to you, not a corporation.”


8. Financial Projections

8.1 Key Assumptions (Year 3)

Metric Conservative Target
Active tutors on protocol 5,000 12,000
Active students on protocol 25,000 60,000
% using Lumen Client (any tier) 40% 55%
% on Pro tier 15% 25%
% using Lumen Relay (paid write) 20% of tutors 30% of tutors
Corporate clients 50 200
Monthly lessons on protocol 150,000 350,000

8.2 Revenue Model (Year 3 — Target Scenario)

Revenue Source Users × ARPU × 12 Annual Revenue
Client Pro subscriptions 15,000 users × $4.99 × 12 $898,200
Relay write access 3,600 tutors × $9.99 × 12 $431,568
AI Co-Pilot (standalone) 5,000 users × $7.99 × 12 $479,400
Corporate dashboards 200 clients × avg $999 × 12 $2,397,600
Dispute resolution 2% of 350K lessons × $15/lesson × 2.5% fee $26,250
Verification service 2,500 new tutors/year × $49 $122,500
White-label clients 30 clients × $499 × 12 + 30 × $1,500 setup $224,640
Total Revenue $4,580,158

8.3 Cost Structure (Year 3)

Cost Category Annual Spend
Engineering (6 FTE) $600,000
Product/Design (2 FTE) $240,000
Relay infrastructure & hosting $180,000
AI/LLM compute $120,000
Marketing & growth (3 FTE + spend) $500,000
BizDev / B2B sales (2 FTE) $260,000
Operations, legal, G&A $200,000
Total OpEx $2,100,000
EBITDA $2,480,158
EBITDA Margin 54%

8.4 Why Unit Economics Are Superior

The open-protocol model fundamentally changes the cost structure:

  • Zero payment processing margin: Lightning fees are <0.1%, not 3%+ for credit cards

  • No customer support for payment disputes: Lightning is final settlement; no chargebacks

  • No churn recovery cost for tutors: Once a tutor is on the protocol, they don’t leave. There’s no competitor protocol to switch to — only different clients.

  • Marketplace growth is not gated by marketing budget: Third-party clients and relays can grow the protocol independently of Lumen’s spend.

  • CAC is structurally lower: The “keep 99%” message is inherently viral among tutors. Student acquisition benefits from lower prices.

8.5 Funding Requirements

Stage Amount Purpose
Pre-Seed $500K–$750K Protocol specification (NIPs), reference client MVP, relay infrastructure, seed tutor cohort
Seed $2M–$3M Full client (mobile + web), AI co-pilot, corporate dashboard, initial marketing, 12-month runway
Series A $6M–$10M Scale relay network, expand AI tools, grow B2B sales team, protocol standardization (NIP advocacy)

9. Competitive Moat in an Open Protocol World

“If the protocol is open, what stops someone from copying you?”**

This is the central question. The answer is multi-layered:

9.1 Moat Components

Layer Moat Defensibility
Protocol Open NIP standards; anyone can build Zero defensibility by design — and that’s the point. The protocol winning is the business winning.
Relay Network Effects Lumen Relay has the most complete, spam-filtered tutoring dataset. Best search quality. Strong. Data gravity. Tutors publish to the relay with the best discoverability.
Client UX Polished, intuitive, AI-integrated client Moderate. UX is replicable, but quality takes years of iteration. First-mover advantage.
AI Co-Pilot Models trained on proprietary session data (with consent), lesson plan templates, prompt engineering Strong. Data network effects: more sessions → better AI → more users.
Corporate Relationships Enterprise sales cycles, SSO integration, compliance features, procurement processes Very strong. B2B is sticky and slow to switch.
Brand & Trust “Lumen” becomes synonymous with “open tutoring marketplace” Moderate to strong. Brand takes years to build.
Contributor Community Open-source client and relay; external contributors improve the ecosystem Strong. Community-owned development out-innovates any single company.

9.2 The Open Protocol Paradox

In a centralized model, you build walls to keep users in. In an open protocol, you build bridges to attract users — and the bridges themselves become the moat.

  • Lumen contributes to Nostr protocol development (NIP proposals, relay standards)

  • Lumen’s code is open-source (AGPL for relay, MIT for client libraries)

  • Third-party developers build on Lumen’s APIs and relay data

  • The protocol becomes the standard for language tutoring

  • Lumen wins because it’s the best-known, most-trusted service provider on the protocol

This is the Linux/Red Hat model applied to marketplaces.


10. Risks & Mitigations

Risk Severity Mitigation
Nostr ecosystem is too small High Don’t rely on existing Nostr users. Build onboarding that abstracts keys and Lightning entirely — users shouldn’t need to know they’re using Nostr.
Key management UX is intimidating High NIP-46 remote signers; “Sign in with Google” flow that provisions a Nostr key behind the scenes; social recovery options. The protocol is Nostr; the UX is frictionless.
Spam and Sybil attacks on open relays Medium Proof-of-work requirements, reputation-weighted relay policies, optional paid relay write access, web-of-trust filtering.
Lightning Network volatility & complexity Medium Auto-convert sats to local currency; abstract Lightning entirely behind a fiat card payment interface. Tutors and students see dollars/euros, not sats.
Regulatory uncertainty (KYC/AML) Medium Fiat on/off-ramp partners handle compliance. The protocol itself is neutral infrastructure — like email.
Centralized competitors adopt Nostr features Low-Medium They cannot. The economic model of 18–33% commission is incompatible with an open protocol. They would cannibalize their own revenue.
Slow adoption by non-crypto-native users High Patience. The internet itself took 15 years. Focus on specific, underserved niches first where the value prop is overwhelming.

11. Roadmap

Phase Timeline Key Milestones
Protocol Design Months 1–3 Publish NIP drafts for tutoring event kinds (30XYZ–36XYZ); launch reference relay; open-source client alpha
Seed Marketplace Months 4–6 Recruit 200 tutors; subsidize 1,000 student trial lessons; iterate on client based on feedback
V1 Client Launch Months 7–9 Public launch of Lumen Client (web + mobile beta); freemium tiers; relay subscription available
AI Co-Pilot Beta Months 10–12 AI lesson planner, pronunciation feedback; Pro tier upsell
Corporate Pilot Months 13–18 10 design partners for B2B dashboard; iterate on enterprise features
Scale Year 2 5,000+ tutors on protocol; mobile app GA; 100+ corporate clients; begin Seed fundraise
Ecosystem Year 3 Third-party clients emerge; NIPs adopted broadly; Lumen becomes the Red Hat of open tutoring

12. The Philosophical Close

The centralized tutoring marketplace model is extractive by necessity — it must take 18–33% of every transaction to fund its own infrastructure, marketing, and profit margins. This creates a ceiling on tutor earnings and a floor on student prices. It locks reputation inside walled gardens and makes the platform itself indispensable.

An open protocol marketplace inverts this entirely.

It says: the protocol belongs to everyone. The data belongs to the users. The relationships belong to the tutors and students. The marketplace is not a company — it is a commons.

Lumen does not seek to own the commons. It seeks to build the best tools for navigating it.

This is both a more ethical model and — paradoxically — a more defensible one. Because when the protocol wins, the best service provider on that protocol wins too. And unlike a centralized competitor, Lumen’s position cannot be disrupted by another company simply offering lower commissions. The protocol already offers near-zero commissions. To compete, you’d have to build a better client, a better relay, a better AI co-pilot — and that competition benefits everyone.

The future of language tutoring is not a better platform. It is no platform at all. It is an open protocol, and the best tools on top of it.

Write a comment