EXTREME DOGFOODING: Skybreach has been moved to a single simple server, using our own home-rolled replacements for EVM indexer, database, API endpoints, search infra.
The de-vercelification of RMRK is now complete.
We have killed Argo, kubernetes, vercel, typesense, elasticsearch, subsquid as dependencies. We kept @filebase and @Alchemy as essential, everything runs on a tiny server now, at a throughput of around 100x what it was 2 years ago, and at 1% the cost.
But we need your help: if you have lands, go try and interact with Skybreach and report bugs you find! The infra we killed was complex and there may be trailing dependencies left behind that will still break things.
There will be bugs on both Singular and Skybreach while we put the final polish on things, but by the end of the week we should be in a mature "let's now build on top of this" state!
Go forth and explore!
We have significantly improved the x402watch directory.
- docs are a part of the UI now, and accessible separately
- we auto-ingest sellers from Bazaar discovery in facilitators
- we check for facilitator gas balance and send alerts if they're low
- we support feature requests and prioritize based on upvotes
- an official Telegram group has been created for bugs, feature requests, clarifications...
Next up:
- historical and current volumes on chains
- seller validity probes
- joint agent-discoverable megabazaar
Something like this would have been impossible just 6 months ago due to the sheer size of the codebase.
Singular, as a behemoth of a project, has always relied on a million dependencies - not just third party ones like externally hosted Postgres and Redis, or an external Elastic and Typesense (yes, all of this!!!), but codebase ones too.
A particular pain point of mine has always been the reliance on NodeJS, React, and Next - three of what I consider to be the most harmful technologies to happen to web dev, ever, and a core part of the Hellstack (https://t.co/nS24SCVTOJ).
Codex helped me remove any touch surface with NodeJS from NX, completely rewriting the whole flow and internals and allowing me to avoid NodeJS completely in both dev and prod (sadly, it's proving a little harder to yank out React and Next).
On this much healthier foundation I then rewrote all graphql, typesense, elastic, and redis queries to use a single well-indexed DB populated by https://t.co/zB33wyAuAv - my high performance EVM indexer.
According to RMRK's grafana, Singular now uses ~75% less RAM, starts and builds 300%-500% faster, loads on all pages 2 to 7 times faster (depending on page) and depends on almost nothing except Alchemy and Kvasyr.
And I've basically only gotten started.
There will probably be bugs, so report them in the https://t.co/Go4yHsADfy group. But we're getting to a nice, context-manageable, infra-stable state, and I'm loving the progress!
@easeev@ChainstackHQ Thanks but the goal is to reduce reliance on third parties and dramatically minimize context (and cost) so I can keep it all in my head.
What took 25 people to run 4 years ago is now just me, and it works at 10% the cost at 10x the perf. And I am just getting started.
AI is mindblowing. The enablement it provides is off the charts and the Codex app is by far my favorite environment (for now).
I created a new high-performance EVM indexer https://t.co/zB33wyAuAv to replace Subsquid and a few Graphql servers / endpoints for RMRK's infra. I didn't originally optimize for @Alchemy CU costs, so I blew the roof off the Pay-as-you-go limit trying to fanatically support 0-conf and reorgs.
I then went on an optimization rampage and got an order of magnitude improvement primarily by using tagged safe blocks and batched log reads. 10x reduction in 2 hours of Pro prompting.
I also made sure the indexer has a user API, and that it has an LLM skillpack that explains how to use this API.
This is where it get super fun!
I then gave this skill to Codex, and told it to rewrite the Emotes app (https://t.co/o9gP2HjaFu), which until now had depended on GraphQL and other shenanigans, to use Kvasyr by making subscriptions to events via the API. All I did was give it an API key to my account on the indexer.
10 minutes later, I had the app completely converted from GraphQL to plain old webhooks with local sqlite, running alongside all other RMRK services on a cheap droplet.
Victory!
good time to announce that I'm not going anywhere until RMRK is more successful than it was at its peak
these contracts will power the "every asset is digitized", but anywhere - that's how well we designed them.
I don't care if my family lives on ramen for a year, I'll do it.
We're aggressively consolidating the RMRK token out of Substrate. The old bridge no longer functions.
Full details here: https://t.co/xRDcWUgiUG
This is part of a larger move to reduce the footprint of RMRK infra and obligations, and to be able to focus on launching revenue generating products.
More info soon.
What this means for you:
- you can still bridge, but the process is a little different.
- the tokens on the Substrate @kusamanetwork side are basically just claims for the Moonbeam / Base version
- we are not "leaving" @MoonbeamNetwork in terms of the token, but we will be phasing out marketplace support. So you can keep the token there if you want, and the @axelar ITS bridge will remain active indefinitely - no rush.
This is all part of the larger consolidation of infra and tokens we mentioned a few times already.
Going forward, as we reduce the chain-support footprint, we will be launching open source tools that allow you to keep collections on those chains active.
That way, even if a chain (and its collections) disappear from Singular, you will have the tools you need to keep them going on your own.
We feel like concentrating attention on the products that are alive and in demand works better than supporting and paying for infra very few people use, in more ways than one.
The first of these mini-products - an open source hyper-performant self hostable renderer for flattening and rendering composable NFTs - is going live this coming week.
Currently live on @base, @0xpolygon, @ethereum, value.eth is a fully decentralized locally-runnable front-end. Access it via value.eth in browsers that support @ensdomains or by hitting https://t.co/GG0FC1fo1N.
The contract is verified on the chain scans (see footer).
Your community VALUES you. How much do you VALUE them?
Introducing value.eth: lock up your @Uniswap LP positions for a time or FOREVER, but keep earning the LP fees.
Reassure your community of rug-proofness regardless of launchpad! 👇🧵
RMRK is under new management.
I'm taking over development from the original team. First action: conviction stacking.
1.5M RMRK (~15% supply) will be permanently locked into LP positions on value.eth - no unlock, ever. Only fees collected.
First 200K already permalocked - check https://t.co/0AoNP0miGN:
- Accumulation stack: https://t.co/ZNQOOsWekY
- Recovery stack: https://t.co/zrR859VZPd
No roadmap. No promises. No chats. Just building.
If you're still holding RMRK, you're along for the ride.
I’ve been in discussions with Steven about the possibility of taking over RMRK and leading its next chapter under my direction. While the original team will not remain involved in any meaningful way, I’d still love to have their input and insights when possible. The project would restart with a very small core team likely consisting of 2 to 3 people originally volunteering, including myself, but I’m committed to growing and strengthening it over time.
I proposed the idea of a community poll, but this still needs to be discussed with the remaining original team. I know there have been a lot of uncertainties surrounding RMRK, and I believe the community deserves more than just another “maybe.” I’ve always been a strong supporter of RMRK, the token, and everything that has been built so far. With some key adjustments, I truly believe RMRK can be brought back to life as a thriving and vibrant ecosystem once again.
Of course, this wouldn’t be a quick or simple process—it would require a huge amount of community support to pull off. But I fully believe in a more transparent, communicative, and collaborative RMRK, where the community plays a bigger role in shaping its future.
The heart of RMRK has always been its community, and I completely understand how some of you may feel frustrated or disheartened by how things have played out over time. I haven’t been as active in the past few years, but after serving as RMRK’s Community Manager, I took some time off and later started a lead generation company that has been fairly successful. Now, I want to dedicate a significant amount of my time to bringing new energy to RMRK. While this is a passion project, and I don’t expect any financial return, I am willing to put in a full-time effort to make this happen—that is—if it’s something the team and community want to move forward with.
I hope to gain the community’s support in shaping this vision. With approval from the remaining team, we can define exactly what this transition would look like, as there are a lot of details to work through. I also believe that if RMRK experiences a successful turnaround, the original team deserves fair residuals in recognition of their past efforts. There are many exciting ideas for events, fundraising initiatives, and new utilities that I’d love to explore in the coming months.
One major step would be raising RMRK tokens for a new treasury, which would be partially locked and used to fund development again. This process would also require some individuals to agree to liquidate their tokens at a fixed rate, likely significantly below market value. Given the current liquidity constraints, this would be a significant risk on my part, but with the community’s support, I believe we can find a path forward, together. There’s many ideas to explore here. But ultimately I would like to see all parts of RMRK succeed once more.
All of this said, it’s important to note that if RMRK sees a significant price spike, this plan could be put on hold, potentially indefinitely. Please DO NOT make any decisions based on this message, as nothing is set in stone yet, there isn’t even an official poll in place. Make decisions based on the existing RMRK and not this one until funding is complete — nothing is final. A dramatic spike prior to fundraising tokens will cause an indefinite delay or inability to utilize the existing token as originally desired. This is the first of many steps that may or may not lead anywhere just yet! So again, please, be smart and don’t base your decisions off of it!
Transparency, leadership, and clear communication would be top priorities in every aspect of this effort. I encourage you to send me a direct message, comment or message me on telegram if you feel you feel you have any value to bring to the table — no matter how big or small.
I appreciate you for taking the time to consider this, and I look forward to hearing your thoughts.
Let’s re-build something great together!
@RmrkApp
The Power of Community in Crypto
1/ The true strength of a crypto project isn’t just its tech—it’s the community behind it. A perfect example? The @RmrkApp community, which refuses to give up, even after the founders left and key devs moved on.