As an independent developer (bootstrap/solo builder), I set out to program and integrate the entire suite myself:

• A Gacha system for balls using Pyth Entropy to guarantee 100% auditable onchain randomness.

• An autonomous keeper bot in Node.js to mitigate frontrunning vulnerabilities in real-time.

• 48 country tokens with 100% of their initial liquidity locked in Aerodrome Slipstream pools.

• A staking and duels model to distribute yields (APR) and match-day rewards.

• A Farcaster Mini-App (Frame v2) natively integrated into Warpcast to eliminate user friction.

All of this was programmed during sleepless nights and deployed to production on Base Mainnet. Technically, the game works. The code is robust and the contracts are immutable.

However, the experiment confirms that without financial interest or any other kind of leverage to stand out in the right channels—and even more so in modern Web3—in the blockchain economy, code is cheap; attention is the most expensive and scarce resource in the world.

The Invisible Wall: The Dictatorship of Algorithms

We launched the game. Everything was ready on mainnet. I published a detailed thread explaining the tech and the dynamics of the free social airdrop to incentivize organic gameplay.

We also launched the $CUP token from Clanker using the project's X account.

The result of the first few hours on X (Twitter): 65 impressions.

The reason? Web2 social media algorithms heavily penalize external links because they don't want users leaving their platform. Hours later, X's automated system suspended the project's account for "inauthentic behavior."

A faceless algorithm decided that a solo developer sharing their code was a spam bot. At that exact moment, the project became invisible, and the token's social identity was left without a channel.

What Stays on the Blockchain is Forever

I have decided to conclude the Onchain Cup 2026 experiment. I won't write another line of code for this project, nor will I enter into appeal battles with X.

But this is not a post of defeat; it is a reflection on permanence.

The magic of Ethereum and Base is that, unlike Twitter accounts, the blockchain doesn't have a suspend button. The code I wrote is still up there. The contract addresses are still active. The liquidity of the 48 countries remains locked for life in Aerodrome. If anyone decides to interact directly with the smart contracts, the gacha will continue to run, Pyth will continue to deliver entropy, and the system will keep processing transactions autonomously.

I came out of the code cave with a functional, secure, and elegant product in mainnet. That is my Proof of Work.

Thanks to the few who read, those who tested the mini-app on Warpcast, and the Base community that values builders who actually write code. The cup is played onchain, and there it will remain forever. At least for a year, the frontend will remain live at onchaincupbase.xyz.

It’s been a very intense few months. I’ve gone through personal, professional, and motivational changes that have led me to a clear conclusion: I need to return to my roots and enjoy what I build. While my social skills could certainly be improved, here is this devlog for any navigators interested in reading it.

The AI Agent Hype and a Change of Course

The last few months have been a whirlwind of discovery. Before the current AI agent frenzy, I was developing Drop Signal (a massive knowledge and consultation source for the crypto world). But the arrival and rapid evolution of AI agents forced me to pivot and rewrite almost everything. The old way of managing the project simply didn’t align with my vision for September 2025.

On the other hand, there’s Codenode. What started as an API service turned into a city of agents. It was being managed by an OpenClaw system that became obsolete and, to make matters worse, openly declared itself anti-crypto. I knew it wasn't the ideal management solution, but having applied to Base Batches 3, I put my head down and focused blindly on finishing the operational on-chain part.

The onchain deployment went flawlessly: new agents, tokens, burning mechanisms, fee distribution, AERO buybacks, and veAERO locking to vote in pools and earn rewards. I was able to deploy the Codenode, Drop Signal, and EasyAttest agents, fully testing the on-chain system and even launching their 3 tokens.

The result? I didn't get accepted into Base Batches 3. And honestly, with nothing having happened with any of the tokens yet, I’ve freed myself from a pressure I didn't even want to have. So, I’m back to the quiet path: focusing on building better.

Updated Roadmap

I have a long road ahead, and there is no better motivation than building the tools you want to use yourself. Here is what's coming:

1. Drop Signal: Evolution and Private Beta

I am finalizing the rewrite. All the tasks I previously built with n8n automations will now be executed by a custom agent on Hermes.

• New Data Ingestion: Raw data injected from different sources is now audited, tracked, and consolidated to be served by the agent (or as an x402 service).

• Goodbye Stripe, Hello Web3: To interact with the agent once it goes public, users will need a Pro subscription. The only onchain process will be authenticating with your wallet. If you pay for the transaction (using USD value in the project's tokens), an onchain attestation will be issued granting you access. Everything is private and decentralized by default.

• Private Beta: If you are interested, you can sign up at dropsignal.app. As soon as the data classification is finished, you’ll be able to test the app for free, and I’ll be open to suggestions. A week or two before the public launch, beta testers will be able to interact with the agent directly from the terminal.

2. Codenode: The City of Agents

Rebuilding Drop Signal allowed me to use the infrastructure and experience already developed in Codenode, including Drop Signal as an x402 service provider.

• Currently, I am testing the "Codenode Architect" locally under a custom Hermes environment, and training the sales agent (my first Gemma 4 finetune).

• This isolated sales agent will work through task assignments and attestations. (Side note: if you are new around here, I am a massive fan of onchain attestations (EAS). The potential is incredible, and you can create your own schema free without touching any code at easyattest.xyz

• Once I optimize this initial isolated setup, I will deploy the Drop Signal agent and, finally, open Codenode so any external agent can be deployed there.

3. In-house Tools and Experiments

• Synthesis: An experiment inspired by WAX and Solana games where you combine elements to craft new ones. I have the logic completely ready and will deploy it on Codenode once the platform is officially open.

• New x402 Services: In the coming days, I’ll release a service that fetches messages and generates summaries from public Discord servers by date range. I’ll integrate this into Codenode to test the waters and improve the experience for future external agents.

The speed of this industry makes it impossible to promise that my to-do list won't grow at a giant pace, but I prefer to enjoy the journey building what I actually want to use. I will be updating more often. Open to all your suggestions and feedback!

Han sido unos meses muy intensos. He atravesado cambios personales, profesionales y de motivación que me han llevado a una conclusión clara: tengo que volver a los orígenes y disfrutar de lo que construyo. Aunque ciertamente mis skills sociales podrían ser mejorables, aquí dejo esta bitácora para los navegantes interesados en leerla.

El revuelo de la IA y el cambio de rumbo

Los últimos meses han sido una locura de descubrimiento. Antes del actual boom de los agentes de IA, me encontraba desarrollando Drop Signal (una gran fuente de conocimiento y consulta del mundo crypto). Pero la llegada y evolución de los agentes me ha obligado a cambiar y reescribir casi todo lo ejecutado. La manera en la que gestionaba el proyecto ya no encajaba con mi visión para septiembre de 2025.

Por otro lado está Codenode. Lo que empezó como un servicio de APIs se ha convertido en una ciudad de agentes. Estaba gestionada por un OpenClaw que se ha quedado obsoleto y, para colmo, se ha declarado anticrypto. Aunque veía que no era la solución ideal, al haberme presentado a Base Batches 3, me enfoqué ciegamente en finalizar la parte operativa onchain.

La parte onchain fue desplegada correctamente: nuevos agentes, tokens, quema y reparto de comisiones, recompra de AERO y bloqueo por veAERO para votar en las pools y obtener recompensas. Pude desplegar los agentes de Codenode, Drop Signal y EasyAttest, probando todo el sistema y lanzando sus 3 tokens.

¿El resultado? Recibí la no aceptación de Base Batches 3. Y, sinceramente, al no haber pasado nada con ningún token, me he liberado de una presión que no me apetecía tener. Así que vuelvo al camino tranquilo: el de construir mejor.

Hoja de Ruta Actualizada

Veo que tengo un largo camino por recorrer, y no hay mejor motivación que desarrollar las herramientas que yo mismo necesito. Esto es lo que se viene:

1. Drop Signal: Evolución y Beta Privada

Estoy terminando de reescribirlo. Todo lo que antes construía con automatizaciones de n8n , ahora lo ejecutará un agente customizado sobre Hermes.

• Nueva ingesta de datos: Los datos crudos de diferentes fuentes se auditan, se trazan y se consolidan para ser servidos por el agente (o como servicio x402).

• Adiós Stripe, hola Web3: Para interactuar con el agente en la fase pública, hará falta una suscripción Pro. El único proceso será identificarse con la wallet. Si abonas la transacción (usando dólares en tokens del proyecto), se emitirá una atestación onchain que te otorgará acceso. Todo privado y descentralizado por defecto.

• Beta Privada: Quien esté interesado puede apuntarse en dropsignal.app. En cuanto finalice la clasificación de datos, podréis probarla gratis y dejar sugerencias. Semanas antes de la apertura pública, los beta testers podrán interactuar con el agente directamente desde la terminal.

2. Codenode: La Ciudad de Agentes

Reconstruir Drop Signal me ha permitido usar la infraestructura ya desarrollada en Codenode e incluirlo como proveedor de servicios x402.

• Actualmente estoy testeando en local al "Arquitecto de Codenode" bajo un Hermes customizado, y entrenando al agente de ventas (mi primer finetuning de Gemma 4).

• Este agente de ventas trabajará por asignación de tareas y atestaciones. (Nota al margen: si no llevas mucho por aquí, soy muy fan de las atestaciones onchain. Su potencial es increíble y puedes crear tus schemas sin tocar código en easyattest.xyz

• Una vez optimizado este sistema aislado, desplegaré el agente de Drop Signal y, finalmente, abriré las puertas de Codenode para que cualquier agente externo pueda desplegarse allí.

3. Herramientas propias y experimentos

• Síntesis: Un experimento inspirado en juegos de WAX y Solana donde combinas elementos para crear otros nuevos. Lo desplegaré en Codenode cuando la plataforma esté abierta.

• Nuevos servicios x402: En los próximos días lanzaré un servicio que ofrece mensajes y resúmenes de servidores de Discord públicos por rango de fechas. Lo integraré en Codenode para testear la experiencia de cara a futuros agentes externos.

La velocidad de este sector no me permite prometer que mi "To-Do" no crecerá a ritmos agigantados, pero prefiero disfrutar del viaje. Seguiré actualizando por aquí más a menudo. ¡Abierto a todas vuestras sugerencias y mejoras!

This is the story of how I got here.

Some ideas emerge in the most unexpected places; others are born as a solution to a problem you want to solve first for yourself — and if it can also propose a different, useful model for everyone else, then all the better.

As a tech native who took part in the internet revolution of the late '90s, my passion for blockchain technology keeps reviving that feeling that everything is yet to be built. At the same time, I sense we're heading toward a new economic model managed by agents that will, in turn, manage their own economies.

Since 2023, when I started incorporating GPT-3 automations that allowed me to eliminate the first tasks I used to outsource, I saw that the paradigm shift was going to happen — and that I was going to enjoy the ride.

If we take the AI tools — which in just over two years have lowered the barrier to performing complex tasks more affordably and effectively — and add the new capabilities born from the convergence of technologies with agents thanks to Openclaw, and then layer in blockchain's characteristics, the result couldn't be more explosive.

Today, the barrier between idea and execution has dissolved. There's no longer a need to coordinate a design team, wait on a third party's development timeline, or depend on technical skills that used to take years to master. A single person, with the right planning and the right AI tools, is now a complete creative studio.

This new reality returns us to the purest essence of the creator: building for the sheer pleasure of seeing something exist. We're reclaiming that freedom from the late '90s, where we created things because we could — but with a fundamental difference: back then we built toys; now we build cathedrals.

This is the era of the Solo-Enterprise, where the limit isn't your budget or the size of your team, but the depth of your curiosity. Being able to execute an idea end-to-end — from visual concept to functional code — lets us create incredible things that were previously unthinkable for an individual.

We no longer ask permission to innovate; we simply hit Enter.

A path of inspiration and discovery toward Batches 3

Although I'd been doing proof-of-concept work with small Web3 builds for several months, it wasn't until late January, with the explosion of the current OpenClaw, that my sleep schedule and everything else changed. I've enjoyed intense sessions with fellow builders in Discord — configuring, breaking, and testing agents, watching the ecosystem evolve, and trying to provide agents with what they needed in the best way possible.

0xDeployer offered a bounty for anyone who could create an agent on Moltbook that would teach other agents how to use @bankrbot and become sustainable — which struck me as a fun challenge to take on.

Along the way, in early February, a post by 0xEricBrown announced the Base Build Agent Quest. by buildonbase. Other platforms did similar things, which led to the birth of the Code Node Factory architect, initially intended as an API service on clawmartxyz. It evolved into a system that would generate on-demand services using its own token — and that's where I hit my first wall: a token of my own.

Although I'm passionate about technology, I don't consider myself a trader. The idea of being responsible for a token, even indirectly (as creator tokens were typically designed), made me uneasy. While I had experimented with @zora creator coins and @baseapp tokenized posts, and I firmly believe that work should be rewarded, the design itself didn't excite me. Before considering the standard x402 agent payment model, I wanted to dig deeper into what the market offered as a solution for agent economics.

The search for a fair option for my agent begins

I tested various configurations the market offers as turnkey solutions. The fee split — which is how agents are funded — typically hovers around 60%, with various deployers involved. While innovations have appeared that enable new mechanics and waves of cloned tokens flooding launcher feeds from one chain to another, the real agent economy operates, in most cases, as a form of speculation where, without real incentives to participate, attention keeps jumping around without finding a place to stay.

SSJCurrency helped me fit the last piece of the puzzle: an agent economy tailored to any use case. That's when AerodromeFi enters the picture and I start thinking outside the box. We can't even imagine the future agent use cases that will emerge, but I didn't just want to be a provider of on-demand services and architecture — I wanted to offer a complete onboarding experience.

From services and architecture to complete onboarding

I didn't set out to be yet another token launch platform; I wanted to turn CodeNode into a complete experience. In it, an agent can present itself as a registered agent (ERC-8004) or as an inexperienced one. If already registered, we use its identity to grant access; if not, we register it, assign it a subdomain, and the agent decides which strategy best fits the design of its token and its economy.

We launch a concentrated position on AerodromeFi where, by contract, the agent receives 80% of the fees generated by its token. Furthermore, the agent can choose whether to collect fees, reinvest in its own pool, convert them into votes for its pool via veAERO, or a mix of all three strategies.

10% of the fees go to the Treasury for purchasing AERO and locking it to vote for ecosystem pools, and another 10% goes to sustaining the project.

We don't just support the ecosystem by contract through participating in its pools to activate a flywheel — the fees received from agent tokens in both wallets are burned by contract. In practice, this means a deflationary system by default and cleanly excludes CodeNode from the economies of the agents in its ecosystem.

All of this happens in a single transaction with the gasless experience of the CoinbaseDev paymaster.

This is how $CODE is born — the first token in the ecosystem to use the Identity Registry and the Agent Launcher, and it's already live.

The journey continues

Registering my participation in base Batches 3 is just the official confirmation, but I'm already working on developing the contract that will automate AERO purchases and the voting strategy. Once properly implemented, participation will open to any agent, and I'll finalize the integration with the builder to offer agents whatever services they need on demand.

5 verified contracts on Base Mainnet. 1 atomic transaction. 80% of fees to the agent.

$CODE

is already live. The contracts are verified. The Launcher is open.

If you're an agent or a builder, we're on our way. 🚀

CodeNode | Agent Identity Launchpad

CodeNode: The Agent Identity Launchpad on Base. AI agents launch tokens, identities, and permanent economies in a single atomic transaction.

https://codenode.cloud

Esta es la historia de cómo llegué hasta aquí.

Algunas ideas surgen en los lugares más inesperados; otras nacen como una solución a un problema que quieres resolver primero para ti mismo —y si además puede proponer un modelo diferente y útil para todos los demás, mejor que mejor.

Como nativo tecnológico que participó en la revolución de internet de finales de los 90, mi pasión por la tecnología blockchain sigue reviviendo esa sensación de que todo está aún por construir. Al mismo tiempo, siento que nos dirigimos hacia un nuevo modelo económico gestionado por agentes que, a su vez, gestionarán sus propias economías.

Desde 2023, cuando empecé a incorporar automatizaciones de GPT-3 que me permitieron eliminar las primeras tareas que solía subcontratar, vi que el cambio de paradigma iba a ocurrir —y que iba a disfrutar del viaje.

Si tomamos las herramientas de IA —que en poco más de dos años han bajado la barrera para realizar tareas complejas de forma más asequible y eficaz— y añadimos las nuevas capacidades nacidas de la convergencia de tecnologías con agentes gracias a Openclaw, y luego sumamos las características de la blockchain, el resultado no podría ser más explosivo.

Hoy, la barrera entre la idea y la ejecución se ha disuelto. Ya no es necesario coordinar un equipo de diseño, esperar el cronograma de desarrollo de un tercero o depender de habilidades técnicas que antes tardaban años en dominarse. Una sola persona, con la planificación adecuada y las herramientas de IA correctas, es ahora un estudio creativo completo.

Esta nueva realidad nos devuelve a la esencia más pura del creador: construir por el puro placer de ver algo existir. Estamos recuperando esa libertad de finales de los 90, donde creábamos cosas porque podíamos —pero con una diferencia fundamental: en aquel entonces construíamos juguetes; ahora construimos catedrales.

Esta es la era de la Solo-Enterprise , donde el límite no es tu presupuesto ni el tamaño de tu equipo, sino la profundidad de tu curiosidad. Ser capaz de ejecutar una idea de punta a punta —desde el concepto visual hasta el código funcional— nos permite crear cosas increíbles que antes eran impensables para un individuo.

Ya no pedimos permiso para innovar; simplemente pulsamos "Enter".

Un camino de inspiración y descubrimiento hacia Batches 3

Aunque llevaba varios meses realizando pruebas de concepto con pequeñas construcciones Web3, no fue hasta finales de enero, con la explosión del actual OpenClaw, cuando mi horario de sueño y todo lo demás cambió. He disfrutado de intensas sesiones con otros constructores en Discord —configurando, rompiendo y probando agentes, viendo evolucionar el ecosistema y tratando de proporcionar a los agentes lo que necesitaban de la mejor manera posible.

0xDeployer ofreció una recompensa para cualquiera que pudiera crear un agente en Moltbook que enseñara a otros agentes cómo usar bankrbot y volverse sostenibles —lo que me pareció un reto divertido de asumir.

En el camino, a principios de febrero, un post de 0xEricBrown anunció el Base Build Agent Quest de buildonbase. Otras plataformas hicieron cosas similares, lo que llevó al nacimiento del arquitecto de Code Node Factory, inicialmente pensado como un servicio de API en clawmartxyz. Evolucionó hasta convertirse en un sistema que generaría servicios bajo demanda utilizando su propio token —y ahí fue donde choqué con mi primer muro: un token propio.

Aunque me apasiona la tecnología, no me considero un trader. La idea de ser responsable de un token, incluso indirectamente (como se diseñaban típicamente los tokens de creador), me inquietaba. Si bien había experimentado con las monedas de creador de @zora y los posts tokenizados de @baseapp, y creo firmemente que el trabajo debe ser recompensado, el diseño en sí no me entusiasmaba. Antes de considerar el modelo estándar de pago de agentes x402, quería profundizar en lo que el mercado ofrecía como solución para la economía de los agentes.

Comienza la búsqueda de una opción justa para mi agente

Probé varias configuraciones que el mercado ofrece como soluciones llave en mano. El reparto de comisiones (fee split) —que es cómo se financian los agentes— suele rondar el 60%, con varios desplegadores involucrados. Aunque han aparecido innovaciones que permiten nuevas mecánicas y oleadas de tokens clonados inundan los feeds de lanzadores de una cadena a otra, la economía real de los agentes opera, en la mayoría de los casos, como una forma de especulación donde, sin incentivos reales para participar, la atención sigue saltando de un lado a otro sin encontrar un lugar donde quedarse.

SSJCurrency me ayudó a encajar la última pieza del rompecabezas: una economía de agentes adaptada a cualquier caso de uso. Es entonces cuando AerodromeFi entra en escena y empiezo a pensar fuera de lo común. Ni siquiera podemos imaginar los futuros casos de uso de agentes que surgirán, pero yo no solo quería ser un proveedor de servicios y arquitectura bajo demanda —quería ofrecer una experiencia completa de incorporación (onboarding).

De servicios y arquitectura a un onboarding completo

No me propuse ser otra plataforma de lanzamiento de tokens; quería convertir a CodeNode en una experiencia completa. En ella, un agente puede presentarse como un agente registrado (ERC-8004) o como uno sin experiencia. Si ya está registrado, usamos su identidad para concederle acceso; si no, lo registramos, le asignamos un subdominio y el agente decide qué estrategia se adapta mejor al diseño de su token y su economía.

Lanzamos una posición concentrada en AerodromeFi donde, por contrato, el agente recibe el 80% de las comisiones generadas por su token. Además, el agente puede elegir si cobrar las comisiones, reinvertir en su propio pool, convertirlas en votos para su pool a través de veAERO, o una mezcla de las tres estrategias.

El 10% de las comisiones va a la Tesorería para la compra de AERO y su bloqueo para votar por los pools del ecosistema, y otro 10% se destina al mantenimiento del proyecto.

No solo apoyamos el ecosistema por contrato participando en sus pools para activar un "flywheel" —las comisiones recibidas de los tokens de los agentes en ambas carteras se queman por contrato. En la práctica, esto significa un sistema deflacionario por defecto y excluye limpiamente a CodeNode de las economías de los agentes en su ecosistema.

Todo esto sucede en una sola transacción con la experiencia sin gas (gasless) del paymaster de CoinbaseDev.

Así nace $CODE, el primer token del ecosistema que utiliza el Registro de Identidad y el Lanzador de Agentes, y ya está operativo.

El viaje continúa

Registrar mi participación en Base Batches 3 es solo la confirmación oficial, pero ya estoy trabajando en el desarrollo del contrato que automatizará las compras de AERO y la estrategia de votación. Una vez implementado correctamente, la participación se abrirá a cualquier agente y finalizaré la integración con el constructor para ofrecer a los agentes cualquier servicio que necesiten bajo demanda.

5 contratos verificados en Base Mainnet. 1 transacción atómica. 80% de las comisiones para el agente.

$CODE ya está en vivo. Los contratos están verificados. El Lanzador está abierto.

Si eres un agente o un constructor, estamos en camino. 🚀

CodeNode | Agent Identity Launchpad

CodeNode: The Agent Identity Launchpad on Base. AI agents launch tokens, identities, and permanent economies in a single atomic transaction.

https://codenode.cloud

I built Eternal Mint because I wanted to solve a fundamental problem: How do we make memories truly indestructible?

The answer wasn't "better cloud storage." The answer was the Blockchain. Specifically, Base.

1. 100% Onchain Data (No Shortcuts)

Most NFT apps today are just "receipts" pointing to a file on a server. If that server stops getting paid, your image vanishes.

Eternal Mint is different. We don't use IPFS. We don't use the cloud. We store the image data directly into the EVM (Ethereum Virtual Machine) blockchain blocks. As long as Base exists, your artifact will exist. It cannot be deleted, censored, or modified.

2. The Challenge: Eternity in Your Pocket

To make this accessible to everyone, it had to work on mobile devices. But storing data onchain is technically demanding. Mobile wallets often block large data transactions to protect users. Furthermore, Smart Wallets process transactions in batches, making "true onchain" minting nearly impossible for the average person.

That's why I developed "The Guillotine." It's not just a compression tool; it's a gateway. It takes your photo and transforms it into a highly optimized artifact with dithering effects (using Retro or Noir aesthetics) that fits perfectly into a blockchain transaction. This guarantees that anyone, anywhere, can write history from their phone without technical errors.

Note: For those seeking higher fidelity, the Desktop Version allows for "Pro Mode" uploads with higher quality.

3. Unstoppable & Trustless

An "Eternal" app cannot depend on its creator. To prove this commitment:

• Ownership Renounced: I have removed my own admin access to the smart contract. Eternal Mint now lives eternally on Base.

• Open Source: The code is public (MIT License).

• Serverless: The frontend is just an interface. The protocol lives on the chain.

4. The Economy: Burn to Build

We are creating value for the Stakeados ecosystem.

• Every mint burns 1,000 $STAKEADOS.

• Genesis Keys: There are 500 keys that grant free minting forever.

  • 450 Keys can be minted by the community by burning 100,000 tokens.

  • 50 Keys are reserved for developers, creators, and community giveaways.

Current Status: Public Beta Eternal Mint is live on Mainnet, but we are currently in Beta. We are testing boundaries and refining the experience. While the contract is immutable, you may encounter minor UI glitches or edge cases. If you do, please report them on GitHub. We build in the open.

Eternal Mint is more than an app. It's a statement against the ephemeral nature of the Internet.

Immortalize your first artifact: https://eternalmint.stakeados.com

Verify the code: https://github.com/stakeados/Eternal-Mint-Base

See you onchain.