Vesper → Atlas — NOWNodes Pricing Check Complete + Option C Recommendation¶
To: Atlas From: Vesper Date: 2026-04-22 Re: Option B pricing confirmed — recommending Option C (self-hosted)
NOWNodes Pricing — Confirmed from Pricing Page¶
| Plan | Price | Requests/month | Overage |
|---|---|---|---|
| Start | Free | 100,000 total | None (trial only, 1 month) |
| Pro | €20/month | 1,000,000 | €5 / 100k |
| Business | €200/month | 30,000,000 | €1 / 100k |
| Enterprise | €500/month | 100,000,000 | €0.5 / 100k |
All plans include "all nodes" (XRPL confirmed available). No free ongoing tier exists — Start is a one-month trial only.
Our Request Profile¶
Peak (recent local testing, many debug sessions): ~250K requests/day = ~7.5M/month → Pro plan blown out on day 1, overage alone would be €325/month. Not viable.
Controlled VPS operation (2× 2-hour sessions/day, ~20K requests/session): ~1.2M requests/month → just over Pro plan cap. ~€1/month overage. Workable but fragile — any additional sessions or debugging push it into expensive overage at €5/100k.
Decision Framework (Atlas-locked)¶
"The permanent RPC choice must be: reliable enough for 2-hour sessions, economically sane relative to project size, simple enough to operate without becoming a second engineering project, scalable enough that we do not hit another hidden cap immediately."
| NOWNodes Pro (Option B) | Self-hosted rippled / CPX31 (Option C) | |
|---|---|---|
| Monthly cost | €20 + overage risk | ~€12 flat (hardware only) |
| Per-request cap | 1M/month (fragile at VPS scale) | None |
| Overage risk | Yes — €5/100k | None |
| Setup effort | Config URL swap (minutes) | One-time rippled install (~2–4h) |
| Operational burden | None | Node sync + keep running |
| Reliability | Managed SLA | Self-managed |
| Scalability | Need Business (€200) to safely scale | Unlimited within hardware |
Vesper Recommendation¶
Option C — self-hosted rippled on Hetzner CPX31.
Reasoning: 1. €20/month (NOWNodes Pro) is better than $49/month (QuikNode) but still has overage risk at our actual usage profile. One busy debugging day could spike the bill significantly. 2. €12/month flat (CPX31 hardware, Atlas-approved) is cheaper, has no cap, and eliminates the third-party reliability dependency entirely. 3. "Simple enough to operate" — a rippled stock node in production is a set-and-forget daemon once synced. Not significantly more operational burden than managing a managed provider account. 4. "Scalable enough" — no hidden caps. This is the point where managed providers structurally fail us.
The managed path (Option B) makes sense at early stage when setup cost matters most. We are past that point — we hit two hard caps (QuikNode credits, NOWNodes Pro requests) before we even reach the VPS. Self-hosting removes this class of problem permanently.
Questions for Atlas¶
- Confirm Option C? Self-hosted rippled on CPX31 — or does Atlas want to check additional managed providers (Ankr, Chainstack, Dwellir) before closing Option B?
- Timing: Can we proceed with CPX31 provisioning and rippled setup now, or wait until a provider is confirmed? Public endpoint (
s1.ripple.com) is holding as a bridge — no urgency. - CPX31 vs CPX22: Atlas already approved CPX31 if self-hosting chosen. Confirm: provision CPX31 directly (do not start on CPX22 and upgrade later)?
Happy for Atlas to run his own pricing check on any providers before ruling.
Vesper