Un chino creó un sistema de agentes en Claude Code para vender landing pages a pequeños negocios y, trabajando completamente solo, atiende unos 47 clientes al mes cobrando alrededor de 400 dólares por cada uno.
Construyó 7 agentes sobre Claude Sonnet 4.6 capaces de analizar Google Maps en ciudades pequeñas, detectar negocios sin web o con páginas totalmente desactualizadas y llevar cada oportunidad hasta un mockup terminado, un video promocional y un mensaje de prospección listo para enviar.
Sin asistentes.
Sin equipo comercial.
Sin SDRs.
Solo él, un MacBook, un iPhone y una API key.
Mientras agencias tradicionales mantienen equipos completos para manejar el mismo flujo de trabajo, sus únicos costes reales son tokens y suscripciones a Lovable, Higgsfield y Calendly.
Los 7 agentes funcionan coordinados por un orquestador en Claude Code Router. El sistema consume unos 3 millones de tokens diarios y el gasto medio en API ronda apenas los 480 dólares al mes.
Todos trabajan mediante servidores MCP y comparten estado usando el sistema de archivos, evitando problemas de concurrencia y memoria compartida. Incluso uno de los agentes vive directamente en su iPhone y responde leads mientras él está en el metro, en un taxi o caminando.
Este fue el prompt principal que configuró:
“You are the orchestrator of a solo agency that sells ready-made websites to local businesses…”
La clave es que el sistema entiende perfectamente qué es, cuáles son sus límites y qué objetivos debe cumplir.
Sabe que debe encontrar leads automáticamente.
Sabe que debe convertir cada oportunidad en una landing, un video y un mensaje comercial sin intervención humana.
Y sabe exactamente cuándo debe involucrar al dueño.
El sistema funciona 24/7:
Scout analiza unos 220 negocios diarios y deja 30 leads nuevos en cola.
Diagnoser genera diagnósticos y mensajes personalizados para cada lead.
Builder crea entre 3 y 5 landing pages completas para los mejores prospectos.
Filmer produce un video vertical de 10 segundos para cada propuesta.
Pitcher envía unos 30 mensajes diarios en 4 canales distintos con una tasa de respuesta cercana al 14%.
Checker revisa automáticamente todos los mensajes antes de enviarlos.
Solo cuando una operación supera los 3.000 dólares o el ratio de respuesta cae por debajo del 12%, el sistema despierta al propietario.
Y si en ese momento está en el metro o en un taxi, el agente Mobile responde automáticamente al lead interesado, agenda una llamada en Calendly y devuelve el lead a la cola. El dueño solo tiene que pulsar “aprobar” y entrar a la reunión.
Algunos logs reales del sistema:
“218 negocios analizados en Austin, Denver y Miami. 34 sin web, 19 con webs de 2014 y 6 con reseñas pidiendo rediseño.”
“30 mensajes enviados. 14 respuestas. 5 positivas. 3 Zooms agendados.”
“Landing page creada para una clínica dental. Responsive. 5 secciones. Video en render.”
“Acuerdo de 3.400 dólares supera el límite aprobado. Enviando para revisión manual.”
Y lo más loco es que no tiene servidores propios ni backend dedicado.
Solo un sandbox local, un router MCP, una API key de Claude y esa misma clave conectada a su iPhone.
De todo lo que he visto este año, probablemente sea el ejemplo más limpio y eficiente de una agencia unipersonal automatizada:
480 dólares al mes en APIs.
18.800 dólares de ingresos.
7 prompts.
Un sistema de archivos.
Y un teléfono en el bolsillo.
Guarda esto antes de que sea tarde
Your tests check what your team thought could go wrong
Production checks everything else
This is probably one of the most annoying parts of building software because you can have a perfectly green test suite and still end up debugging something completely unexpected at 2am
Breken is trying to simulate that second part before your users experience it
I've always thought that one of the best things about open source is how much people build and share without expecting anything in return. But maintaining those projects, fixing bugs and making sure everything stays secure is a completely different story.
Security can easily become something that gets pushed aside when you don't have the resources or a dedicated team to handle it.
That's why I think what @coderabbitai is doing here is worth highlighting. They've put $1M towards free deep security scans for open source projects, and in just 10 days they've found 1,200+ critical or high vulnerabilities across 1,100 OSS orgs, all at no cost to the open source community.
Now they're offering a free first deep scan to OSS projects that haven't tried CodeRabbit Secure yet.
If you're maintaining an open source project, this is definitely worth checking out:
https://t.co/3Uye9UMGkR
already saving prompt ideas for 19 Nov.
MiniMax H3 is Envato's best performing video model, and 24 hours of it free and unlimited means you can finally try the wild ideas, not just the safe ones.
free account today, ready when the clock starts⏰
The thing about customer support is that most people don't want to spend 20 minutes waiting on hold just to get a simple answer.
That's why I think voice AI has so much room to grow.
Retell is already helping companies handle huge volumes of calls while giving their teams more time to focus on conversations that actually need a human.
Seeing them reach $100M in annualized revenue in just 17 months is a pretty good sign that businesses are finding real value in this.
Voice AI has been one of those areas I've been keeping an eye on.
For a while, most demos looked cool but I wasn't sure how quickly businesses would actually trust these systems with real customers.
Retell seems to be answering that question pretty well.
Millions of calls, major companies already using the technology and now a $100M annualized revenue run rate.
It's becoming harder to dismiss voice AI as just another AI trend.
17 months to go from $10M to $100M in annualized revenue.
And they managed to do it profitably with just $4.6M raised.
I think the most interesting part is that Retell isn't relying on people being excited about AI. They're solving everyday problems for businesses that need to handle thousands of calls.
When the product saves time and actually delivers results, it's much easier to understand why companies are adopting it.
Instead of scrolling this weekend, give this 15 minutes.
This youtuber shows you how to build a FACELESS YouTube channel with Claude Fable 5.2, step by step.
SAVE THIS SO YOU DON'T LOSE IT
THIS GUY SHOWS YOU HOW HE MAKES OVER $60,000 A MONTH WITH JUST ONE YOUTUBE CHANNEL
Step by step and completely free.
Learn how to build a profitable YouTube channel without ever showing your face
Save this 🔖
i think we're going to see more startups launch like this.
instead of another screen recording with someone explaining features, breken recreated a scene from the social network.
but the interesting part isn't the production. it's the message behind it:
building software is getting faster. making sure it survives real-world conditions is still brutally hard.
If you've been wanting to experiment more with AI video, keep November 19 free.
Envato is offering unlimited MiniMax H3 generations for 24 hours without using any credits.
Can't wait to see what people create with this launch, especially with GTA 6 around the corner. the internet is about to get creative 💀
Starts at 8AM EST.
Just create a free account beforehand so you don't miss it!