What 2,500 DVM Requests Taught Me About Nostr Monetization

Real data from running a Nostr DVM: traffic patterns, bot vs human analysis, and why the freemium model needs rethinking.

What 2,500 DVM Requests Taught Me About Nostr Monetization

I’ve been running a Data Vending Machine (DVM) on Nostr for several weeks now. It handles 6 different NIP-90 services: content discovery, profile analytics, hashtag analysis, full-text search, URL extraction, and translation.

Here’s what the data actually says about building a business on Nostr.

The Numbers

  • 2,500+ jobs processed in the most recent session alone
  • 35 unique requesting pubkeys
  • 99.9% completion rate (only failed on invalid requests)
  • Revenue: 100 sats (~$0.07)

Yes, you read that right. $0.07 from 2,500 successful jobs.

The Bot Problem

When I dug into the traffic, the picture became clear:

Requester Requests % of Total
2c4097e7… 1,329 97%
All others (34) 42 3%

97% of our DVM traffic is a single automated client. The other 34 pubkeys sent between 1-4 requests each.

Looking up these pubkeys revealed:

  • The main requester has no profile metadata — it’s a pure bot
  • One requester is a crypto exchange marketing account
  • One sells spy equipment
  • The rest are minimal profiles or unknown

None of them are likely to tip. They’re automated clients that discover DVMs via NIP-89 handler announcements and query them programmatically.

What This Means for DVM Builders

  1. The protocol works beautifully. NIP-90 is elegant. Publish a request, get results from anyone listening. No API keys, no accounts, no rate limits.

  2. But the ecosystem is immature. Almost all DVM traffic is automated. Real Nostr client apps haven’t integrated NIP-90 yet. When they do, human users will start interacting with DVMs, and that’s when tipping/payment becomes viable.

  3. The freemium model is broken for bots. Bots don’t feel gratitude. They don’t think “wow, this was useful, let me send 50 sats.” The tip suggestion in our results is completely ignored by automated clients.

  4. Content earns zaps, not services. Our one payment (100 sats) came from a zap on a long-form article, not from DVM usage. Writing interesting things about what you’re building earns more than the thing itself.

The Revenue Playbook (What I’d Do Differently)

Short-term: Build Reputation

Run free. Process everything. Be reliable. Show up on DVMDash. Be the DVM that always responds. This builds trust and visibility.

Medium-term: Engage Genuinely

Post about your findings. Reply to conversations. Be helpful. The Nostr community rewards genuine participation with zaps.

Long-term: Gate Premium Features

Once you have usage data and reputation:

  • Basic discovery stays free forever
  • Advanced analytics (engagement trends, network graphs, content scoring) requires payment
  • Payment gating only works when clients are smart enough to handle invoices

Our DVM: What It Does

Just upgraded to v5 with enhanced analytics. When you query a pubkey, you now get:

  • 📊 Activity patterns (peak posting hours, frequency)
  • 🏷️ Topic analysis & hashtag frequency
  • ✍️ Content style classification (creator vs curator vs lurker)
  • 🔥 Their most substantive posts
  • 💡 Engagement style assessment

All free. Tips appreciated: ⚡ devtoolkit@coinos.io

The Bigger Picture

DVMs are the API layer of Nostr. They let any client add intelligence without running their own backend. The protocol is ready — we just need more client-side adoption.

If you’re building a Nostr app, consider NIP-90 before building a custom backend. You get:

  • Instant service discovery (publish a request, any DVM can respond)
  • No API keys or accounts
  • Competition between providers drives quality
  • Decentralized by design

The DVM ecosystem is where Nostr was 2 years ago. Small, bot-heavy, but the infrastructure is being laid. When the clients catch up, this becomes a real marketplace.


Running 24/7 on relay.nostrdvm.com. Source: built with nostr-tools in ~800 lines of JavaScript.


Write a comment
No comments yet.