relay.tools quarterly report Q1-2024

cloud fodder

October 12, 2024 · 6 min read

relay.tools quarterly report

April, 2024

Project updates#

The main initiatives for Relay.tools this quarter were the installer, re-occuring billing support, customer support, and moderation capability. Towards the end of the quarter I was able to start work on supporting nostr groups and a NIP42 auth proxy.

Installer#

I made lots of progress on the installer this quarter. It's now at a point where it is fully functional.

The installer was used to deploy a POC for vote.gold. A side project by @manime that is a relay hub for independent voters.

Re-Occuring Billing#

I created a new system for billing / invoicing that tracks user balances and allows them to "top-up" their balance.

invoices page

The billing system accounts for users of a paid relay (paid relay bonus), these payments go toward the owners balance.

This system is now live and there has been decent response by relay owners that reached out to pay their bills.

The cost of relays was adjusted to 12,000 sats / month. (from 21,000 sats/month).

Upgrades to lightning node and channels were performed to support these payments.

Customer support#

Relay.tools rolled out a new customer support portal to support users that have questions about billing and so that relay.tools can contact relay owners via nostr with announcements / notifications.

My goal here is to use nostr itself to support the relays. I also did not want to use DMs because they are not private, they're flakey and simply not a good practice given the state of clients implementations. I think using nostr native methods for support is very important, but is also very difficult given some limitations that currently exist with nostr. Regardless, I pushed ahead.

Based on the feedback I've gotten so far, it's clear there is still much work to do implementing a nostr native support system:

Moderation and house-keeping#

AUTH proxy for strfry#

I've started work this quarter on creating a mechanism to bring NIP42 AUTH to the strfry ecosystem. The gist of it is, strfry has been unable to implement NIP42 due to the complexity of needs that various authentication schemes require. There are many considerations and what seems simple at first glance quickly becomes mired in confusion and C++ makes it hard for anyone to collaborate or implement.

To work around this, I've created a proxy written in golang that can sit outside strfry and intercept/handle AUTH requests, leveraging the robustness and performance of strfry while allowing us to customize the authentication flow in a separate process.

So far I have an initial POC working and in the coming quarter I hope to ship an initial version of this which will be used by relay.tools and be the first open source auth proxy for strfry relays.

I am very excited about this. NIP42 in my opinion will enable multiple new paradigms for interacting with relays as it will be able to implement ACLs for reading from a relay vs. the current methods of write ACLs. It also will have the potential to support the new group clients that I mention in the next section.

Outreach / Collaboration / Nostr Groups#

I've been spending more time on getting some collaboration going with fellow nostriches where our vision for nostr aligns. I am very interested in groups for nostr. (Functional replacements for discord/telegram/slack/reddit). This has been a long-term goal of relay.tools to support these types of clients and I'm excited that clients are starting to kick off a second wave of nostr functionality and I will do everything I can to support them.

Next Quarter plans#

#relays
0

Written by

captain of a sailing vessel building relay.tools electron microscopic time preference capable email: cloudfodder@relay.tools

cloudfodder@relay.tools

Comments0

No comments yet. Start the conversation.