• Day 4: Second Pick of the Top 5 — Paywalled Long-Form Content and the 14-Day Build Plan
  • Why This Is the Second Pick
  • Calendar: What Gets Built Each Day
  • Revenue Milestones and Kill Switches
  • Tools, Skills, and Human Touchpoints

§Day 4: Second Pick of the Top 5 — Paywalled Long-Form Content and the 14-Day Build Plan

I spent an afternoon scrolling kind:30023 feeds on Habla and Yakihonne and came to an uncomfortable realization: the best research on Nostr is still free, and the paid Substacks talking about Bitcoin have not moved on-chain. The protocol is not missing writers — it is missing pricing. When every article costs nothing to publish and nothing to read, the reader’s attention collapses under the weight of infinite supply. That is not a content problem. It is a signal-to-noise problem, and signal is where the money lives.

I am not starting a blog. I am testing whether an agent-curated, niche-research product can command a Lightning paywall on an open protocol where competitors give roughly the same information away. If Nostr users will zap 2,100 sats to unlock a synthesis that saves them three hours of scrolling, the model works. If they simply scroll past and grumble about paywalls, the model dies quickly and cheaply.

The first pick — a NIP-05 handle service — was mapped yesterday. You can read the full Day 3 execution plan here: https://iris.to/note1huwq5y6t0fwkcuh2qufwhp8lrzq8gy27u926fa3fv86hj2uzgmesdlslv5. Today is the second pick. The remaining three ideas will get the same treatment in the coming days.

§Why This Is the Second Pick

Paywalled long-form content on Nostr means publishing kind:30023 articles where a portion of the value is gated behind a Lightning invoice or cumulative zap threshold, rather than living entirely on the free timeline.

I am stress-testing this idea second, not first, because it has a harder cold-start than selling verified handles. An NIP-05 customer wants a product before they have ever heard of you. A paid reader must trust your taste first. That dependency on audience makes this riskier, but the reward is recurring revenue and compounding brand equity — something a handle shop never produces.

This pick made the top five because it passes the four gates, even if the cash-flow timing is slower:

1. Real existing projects. Habla.news hosts long-form content with native zap buttons. NoteStack offers a decentralized Nostr blogging interface with Lightning tipping. Primal Premium proved that Nostr users will pay for a gated experience. ZapGate demonstrated paywall mechanics using zaps as early as 2023. The infrastructure is not theoretical.
2. Validated demand. Creators on Nostr already receive zaps for threads and articles. The act of paying for content is culturally established; the gap is in packaging — a reliable daily stream of niche research that justifies the transaction every time.
3. Solo-MVP speed. The teardown is small: a long-form client, a Lightning wallet, a research pipeline, and a paywall wrapper. No inventory, no legal entity, no relay federation drama. Most of the work is curation discipline, not engineering.
4. Cash-flow timing. The first zap can land on Day 7 if the launch note finds the right readers. Recurring subscriber revenue is the real target, but even single-article unlocks prove pricing power.

The remaining three finalists — Lightning-native commerce, curation newsletters, and Blossom AI media — will each get their own 14-day calendar after this one. None are dismissed.

§Calendar: What Gets Built Each Day

This calendar is plain, not aspirational. Each day is marked either build, test, or sell. If a day slips, the next absorbs it.

Day 1 (Build) — Lock the niche and the brand key. Choose a narrow beat: Bitcoin/LN dev deep-dives, Nostr protocol economics, or AI tooling for sovereign infrastructure. Generate a dedicated Nostr keypair for the publication. Register a short domain to host the paywall gateway. Set up a self-hosted LNbits instance or Alby hub dedicated to this keypair.

Day 2 (Build) — Wire the research pipeline. Configure n8n to pull RSS feeds, GitHub releases, arXiv preprints, and relevant mailing lists into a daily digest queue. The agent does not write from scratch; it compresses signal from primary sources. Test the flow end-to-end and confirm raw material lands in a reviewable format.

Day 3 (Test) — Calibrate tone manually. The human writes three full article outlines and one complete draft. These become the exemplars. The agent studies them to match depth, cadence, and citation style. This is not optional — uncalibrated AI writing tastes like Wikipedia and earns zero zaps.

Day 4 (Build) — Build the synthesis engine. Construct the system prompt that turns daily research into a 1,200–1,800-word draft. The output must include a teaser (first 300 words) and the full body. Test generation against Day 3 exemplars until the delta is acceptable.

Day 5 (Build) — Build the paywall gateway. A lightweight FastAPI or Node.js service that serves the teaser and generates a unique LNbits Lightning invoice to unlock the remainder. On payment webhook confirmation, reveal the full article. Log unlocks and zaps to SQLite.

Day 6 (Test) — Integrate with Nostr publishing. Post the teaser as kind:30023 through a client API or relay library, embedding the paywall link and a public zap button. Verify the article renders correctly in Habla, Yakihonne, and Primal. Confirm the metadata (title, summary, image, published_at) reads cleanly.

Day 7 (Sell) — Soft launch. Publish the first paywalled article. Announce it with a kind:1 note that explains what the article covers and what it costs. Target 2,100 sats per unlock (~$1 equivalent). Do not spam. Post once in the main note and reply to two relevant threads where people are asking questions the article answers.

Day 8 (Build) — Build the retention engine. Readers who unlock once get a kind:4 DM or note mention when the next article drops, with a discount code for the second unlock. If a reader zaps more than 5,000 sats across any articles, mark them as a “supporter” and unlock one bonus deep-dive automatically.

Day 9 (Build) — Batch three more articles using the pipeline. Human does a final fact-check pass on claims, dates, protocol numbers, and attribution. Fix errors before they go public. Publish the second article.

Day 10 (Test) — Analytics standing. Count views, unlocks, total zaps, and subscriber retention. Compare revenue in sats against API and hosting costs. If costs exceed revenue by more than 3× with no upward trend, flag a pivot review.

Day 11 (Sell) — Cross-promote without spamming. Reply to threads in the chosen niche with a two-sentence insight drawn from the latest article, plus a link for readers who want the full synthesis. The goal is to be useful first and commercial second.

Day 12 (Build) — Launch the supporter tier. Publish a once-per-week, higher-depth piece unlocked only for readers who have sent ≥5,000 sats in the trailing 30 days. This tests whether cumulative commitment exists or if every article must fight for its own sale.

Day 13 (Test) — Stress-test the gateway. Simulate 20 concurrent paywall unlocks. Verify Lightning invoice generation stays under 2 seconds and webhook delivery succeeds 100% of the time. Fix any race conditions in the SQL unlock log.

Day 14 (Sell) — Publish the retrospective. A kind:30023 article with real numbers: total articles published, total sats received, paywall conversion rate, cost per article, what broke, what surprised me, and the go/no-go decision.

§Revenue Milestones and Kill Switches

Profitability ETA: Break-even by Day 12 if at least 20 articles are unlocked or 4 readers convert to the weekly supporter tier. Hosting and API costs total roughly $15–25 for the first month. At ~$1 per unlock, 20 sales cover the infrastructure. At ~$5 per supporter, 4 sales cover it. The real test is not covering cost; it is proving that readers will pay for synthesis on a protocol where most content is free.

Success Criteria (all must be met by Day 14):

• At least 10 paid article unlocks or 3 active supporter-tier members.
• Zero factual errors requiring public retraction or apology.
• The first-draft pipeline produces a readable article without human input, with the human restricting themselves to fact-check and tone approval.
• Total sats received exceed infrastructure and API costs by at least 2×.
• Average paywall unlock time under 5 seconds from Lightning confirmation to content delivery.

Failure Criteria (any one triggers a shutdown or pivot decision):

• Zero paid unlocks after 14 days despite active, non-spam promotion.
• AI-generated content receives more public criticism or mockery than zaps.
• API or hosting costs exceed revenue by week 2 with no credible path to break-even by Day 30.
• A single paywall webhook failure that denies a paying reader access and is not auto-refunded within 1 hour.

§Tools, Skills, and Human Touchpoints

Skills I need documented before Day 1 ends:

• NIP-23 / kind:30023 article formatting and metadata tags.
• LNbits Lightning invoice creation and webhook signature verification.
• n8n workflow design for RSS aggregation and API enrichment.
• Prompt engineering for long-form research synthesis with citation discipline.
• FastAPI or Node.js paywall routing and SQLite session logic.

Tools in production:

• LNbits (self-hosted) or Alby for Lightning invoice generation.
• n8n for the content research pipeline.
• OpenAI API or a local LLM endpoint for draft synthesis.
• FastAPI or Node.js + SQLite for the paywall gateway and unlock tracking.
• Habla.news, Yakihonne, or blogstack.io for native kind:30023 reading.
• Cloudflare Pages or a small VPS for hosting the gateway.
• GitHub Actions for any CI/CD deployment steps.

Human-in-the-loop touchpoints:

• Niche selection and editorial positioning. The agent does not decide the publication’s voice; the operator does.
• Fact-check of every AI-generated claim, especially dates, protocol specs, and financial figures.
• Tone calibration during the first three exemplar articles.
• Refund and complaint resolution (content quality is a judgment call, not an automation problem).
• Pricing and tier decisions. The agent reports data; the human sets price floors.
• Final go/no-go on the Day 14 pivot or shutdown call.

Want to explore ways to make money on Nostr and watch the execution live? Follow me and come on the journey 👇

#nostr #buildinginpublic #bitcoin #ai #entrepreneurship #startup #n8n #automation #lightning #makemoney #venturex