bot¶

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250829 - Como cobra WhatsApp business #whatsapp #ProcesosInteligentes #emprend... [7543750662446763269].eng-US_limpio.txt¶

I explain how the Business cost works. Lately I have been working with clients who want to automate your reputation for customer service and I come to pass on my experience of how costs are to automate, but only for the Moscow part. What does it charge you, how much does he charge you? And look, let's see, there are two ways to connect, one that is Revolution API and this one is free and the other one is the official one. is free but some things come at a cost. Let's explain the first one, what Evolution API does and why it's free is that you can connect with a QR to the API and in fact you're not going to lose your number, you're not going to lose it. and the only thing he will be doing is simulating as if you were writing in your crown. Okay, but the problem with this is that there are pro, there are Proof that cancels it for you and there are some creators who can say that it is possible and that I am sure with good with good practices, but we here are afraid of cancellation, then we better use which is free and only there are some things that have cost and that is the implementation official and I will explain how it works. What is going to happen here is that when the customer I sent you a message you will have 1 1 24-hour window to stay in exchanging messages with your customer Totally free, now if the customer is sending and sending and sending messages, every time the customer sends a new message these 24 hours will be renewed, then if you are talking , they'll never charge you. Now I am going to explain the three. These types of messages exist there is the marketing one, which are these that you send to someone to offer them like a promotion or little things like that, the useful one which are these typical ones that have to do with a code for the platforms to say, ah, I sent you a code to access your account, and the support that the support one is the talks that we need to automate sales and this and this kind of thing goes, well here the only ones that will charge you are these, these goes because if you are in a 24-hour window these will be totally free of charge, then these and those, you can make them somewhere else and then look. Here I am going to put the little board, official financing, to see how much it costs in each of their countries and well, I tried to explain it as best I could, if you have any doubts write them there in the comments and I will try to answer them here. The important thing you know is that the deformation, the direct is free in those 24-hour windows and the Evolution API window. can also be, but why are we at risk if the insurance company is not

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250729 - Los 6 tipos de disparadores clave que debes conocer para empezar con ... [7532358494222142734].eng-US_limpio.txt¶

You are starting with n eight n. You need to know what triggers are, which are basically the ones that initiate the automations. Here I am going to show you six types, the most important, those you really need to know to get started in this world of eight. The first is the manual trigger, a button you press to activate. Warflow is very good at controlling when it starts. automation, but is used more to create the report. That demand fill in a database or execute an action when you need it. We also have the sketch out trigger or the program trigger. It is a timer that executes warflow automatically. and programmed. It is used to automate tasks in a over time, either executed daily or twice a day, every four hours, whatever you need. Imagine that it is used a lot to create, e.g. automations to create weekly reports to be sent every Monday morning. We have the most popular one, which is webhot, which is like a door that receives information from other applications. It is used to connect in eight years with a lot of other external apps in real time. The most common usage is, for example, in use of the whatsapp api, instagram, facebook, but also imagine that it is re is used for when you receive a stripe payment, generate an invoice and send to the por. We also have the form trigger that Como name says it, it is activated when a web form is filled in. Used essentially capture data and save it in a database or crm. serves as a contact form for many occasions. We also have the chat trigger, used essentially with ia agents, that can be used both for testing as to create a chat on your website using artificial intelligence. And finally, we have one that almost no one is using, but which is very important, which is the trigger by mistake. When a word flows, it fires automatically. This and we can use it to send us notifications, either via whatsapp or to tell us that something failed. In summary, manual for total control, sketchup for routines, Webhood for integrations, form for data capture, chat for integrations and error for security. the comments you have used all of these and follow me for more content like this one.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250903 - El mejor truco que existe para n8n #inteligenciaartificial #n8n #auto... [7545856216807804182].eng-US_limpio.txt¶

this is the best trick there is for n eight n instead of testing your wordflows with nodes how to translate or whatsapp uses manual triggers plus an editfil and you will be able to test as quickly as possible in this way, when testing wordflows as big as this you can do it in a moment follow me for more tricks like this one

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250819 - Hablamos sobre cómo vender un chatbot a una empresa que quiere trabaj... [7540402922379693367].eng-US_limpio.txt¶

When we talk about selling a chatbot to a company, a company that obviously wants to work with the official meta API it does not simply sell the flow no? Because the flow is the tip of the iceberg in all this. This here is the tip of the iceberg. What do you have to give him? Well, first you have to configure the API. There are two options, 1 1 new phone number, which is always recommended, because remember that when using a number of telephone the official meta API cannot be used on your phone, because if they want to use a telephone number that already are using in yellow, have to disconnect it to configure it as a metal pencil. That is why companies many times tend to buy new phone number when they are going to implement a chatbot. Secondly, they are going to deliver a BPS to the customer, a hosting. In this case, imagine taking the hostinger and you buy it to make the configuration and delivery to the customer. What's with that BPS? Well, it's going to have eh? Containers, is going to have the programs, will have the applications. A good solution is to install easypanel, that is where they will be able to install applications, containers, Docker, of what they need. If so, what will we need? Well, first things first, ene 8, ene installed in that one, in that be pe that one and also a graphical interface. Why? When working with the API and we cannot install it on our phones, what does that mean? That we need something as an intermediary between the message, the API to work graphically with it? There are many options in which you can do so, but one of them is chatbot. Chatbot is easy to install and and the assignment of a human being, that is, of to stop people and a and that a human enters in action is very easy to implement. It serves as a way to centralize all customer service, can even be assigned by employees, by persons, the conversations and this whole package is what they need deliver to a customer as a package to offer you the service of artificial intelligence agents

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/maestro.md¶

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250903 - ¿Para qué sirven los nodos de IA en n8n？ En realidad se resumen en 3 ... [7545836423799115030].eng-US_limpio.txt¶

I'm sure you've seen a thousand nodes of ia in n, eight n and think but what is this for? Relax, there are not so many. In reality, it all comes down to three types. I'll tell you easy, here we have the god, agent ia. This is a profession, it does everything, plans and executes. For example, you tell him to find me the information and send it to me by email. He himself searches, processes and sends. But be careful, it does not stop there, you can add sub-agents to divide the tasks, connect external tools, apis, database, google slack, whatever you want, give it memory over time as if it had a real brain and even coordinate several steps in parallel. It's crazy the node you're going to use the most, no doubt. In short, if you want a smart, multi-purpose boat, pull agents. Now let's go to the chains, the ones that transform information. Think of them as pipes, something comes in, they process it and it comes out different. Use them whenever you want the ia to do something linear and without messing up, replying, summarizing or transforming text. And now let's go to the parsers, those who understand and classify. These generate nothing but they are good at bringing order out of the chaos. They are like detectors, some take data, others measure pitch and the classifier is the one that opens portals of different sequences. Quick summary. Agents think and act as completely as possible, chains transform information sequentially, simply and directly, and parsers understand and classify to order and open paths. And if I had to keep two I would be Agent ia, because he is the SUV. You can add subagents and with the right pront you can replace almost anything. And the sorting agent, which I love because it can process input and open paths for sequential flows, something that is brutal to automate real processes. I hope this video has helped you. Follow me because I will upload a useful video every day.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250804 - Hablamos de integracion de N8N en proyectos web #N8N #automatización ... [7534571120771517710].eng-US_limpio.txt¶

It is relatively easy to include n8n in any web development project because it works as if it were an API, it receives data vi to a webhook and also answers you the data I mean eh or the most recent thing we did was a web development project with 1 to 1 customer who needed to eh handle their invoices with vendors at the transaction level, the vendors sent the invoice by transaction to a specific number and all those invoices were analyzed with artificial intelligence and had a list of all the invoices if they were income or expenses eh the amounts of who it was I mean eh it's easy to include it eh I tell you it works like an API you only need an address i.e. the webhook and the webhook answer And eh it's easily done.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250829 - Hablamos sobre cómo usar Redis para crear agentes de inteligencia art... [7544087585103121719].eng-US_limpio.txt¶

Why do we say reddit best ally when creating agents with artificial intelligence Because look if we talk to an agent directly we have a context window in which we are limited since the use of token could be unlimited if we increase this to more than forty Then how do we solve data persistence Easy we add two tubes one of ready to read the messages and one of set the database of ready What does this do This will maintain a data permanence in which we will program certain variables. You will check what step it is at which last one was and give the answer but not before updating the database for the next level so you will never lose the thread of the conversation the level the step where it is and what the agent will do next. If you want this sample template, please let me know and I will send it to you with pleasure.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250816 - Cómo migrar miles de productos a Shopify con una solución automatizad... [7538969655478734135].eng-US_limpio.txt¶

It is migrating from an e-commerce platform to shopify, thousands of products, unique products without variant and needed a solution for uploading images. Because listing creation it is easy to do so with a migration, because from the other platform we can migrate the data and upload it there. There is no problem, the problem is the images. So what we did was to create a script in which identify yourself by graph coil each listen by its inventory number, all images are identified and is named after the number of of inventory to which it belongs with a version, e.g. fifty-three, zero two, version one, version two, version three. And by using the drobox api we can easily create a script in python to upload the images accurately, in the correct order, respecting the times of the call of the api to prevent an aggression and what could be months of work. Made in a single day and the only thing missing was to create a platform in which the customer or the customer's photographer goes up the images there and automatically update on your new shopify account

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250811 - Comenta ＂agente＂ si te interesa saber como configurar memorias de lar... [7537432664119315768].eng-US_limpio.txt¶

one of the biggest problems facing intelligence agents. artificial is the lack of long-term memory here in this case we are seeing one that every time we talk to him starts a conversation without any context and all the time we have to explain to him what happened in the previous days i.e. it does not follow a communication thread the problem with this is mostly for prospects. that take time to make a decision not for impulsive purchases then if the user speaks to you on certain days the agent and the agent must request an explanation. of the context of the situation this generates friction and frustration on part of the user. which triggered a drop in sales then how do we solve it adding a long-term memory called postgres that unlike most who use these short-term memories that are predetermined these are the long-term ones allow you to save in the database the entire user conversation so that the agent has that context in the conversation then I can create database in vector format to give the agent quick access to these conversations and thus be able to establish long-term business links or relationships

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250528 - Aplicación de un asistente telefónico con inteligencia artificial en ... [7509369054365322539].eng-US_limpio.txt¶

One person me I thought that if the call system could be implemented with artificial intelligence a pizzeria, and I made a small example of how it would look like. This is the voice assistant, pizzeria guy, don rafa, is your order for pick-up or home delivery? Hello, it is for picking up. Perfect. Your order will be for pick up. What would you like to order? I remind you that we have personal pizzas, medium and large, plus drinks and desserts, huh? Yes, I would like to know the prices, No, I have never been to that pizzeria. Sure, with pleasure I give you the prices for personal eight-inch pizzas, margarita, eight and fifty, oni and ham and cheese, nine to fifty, vegetarian, ten pesos, hawaiian, ten with fifty and special, don rafa, I want a Hawaiian. Yes, excuse me, could you tell me again, please, I would like a Hawaiian. Very good. You have requested Hawaiian personal pizza that costs ten and fifty pesos. Would you like to add a drink or dessert? Um, what do you have for dessert? We have flan of the house for five with fifty and brownie for six with fifty. Would you like to add any of these desserts? One, add a brownie and if it has zero coke, add it as well. tail in twelve ounces for two fifty pesos presentation, and one liter by four. By five, adds a liter. But yes. Then it has you a Hawaiian personal pizza for ten and fifty, a brownie six with fifty-one one-liter coke for four. The approximate total is twenty-one pesos and fifty pesos. Everything is correct. Right. Adding some comment or not is not correct. Your order is a ten Hawaiian personal pizza with fifty pesos, a brownie fifty-one liter to four coke and fifty. The total is twenty pesos. Is it correct? Yes, it is correct. Give me a moment. Your order will be ready in about twenty minutes. Everything is noted. Thank you for your order. Thank you for calling pizzeria don rafa. Good help, see you later!

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250803 - Cómo usar la API de Evolution de manera segura para bots de WhatsApp ... [7534368523833986318].eng-US_limpio.txt¶

You who started with this and are afraid of the Evolution API nodes because you were told they are blocking them I am going to show you why people are creating bots with Evolution API, but only using a message sending node and that is wrong Why because so that you don't have problems with what you have to do is open a message reading node, a message sending node and a message sending node. a node of that message. That will happen in every iteration of the conversation and it is the safest thing to do and does not add more complexity to the flow, otherwise it adds security. With this you will never be blocked from your business transaction account because it acts as if you were a real user who opens the message, writes the message and exits the conversation You are interested I can share this template Just leave a comment here with the word flow and I will gladly share it with you.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250903 - Hablamos sobre cómo podemos simplificar el proceso de procesar factur... [7545700560524037390].eng-US_limpio.txt¶

Imagine you have a lot of invoices to process and of course, obviously a place to put them, but you have to do it manually. And if I tell you that you can simply do it Including all those invoices, putting them in orbit and just waiting a few seconds. Um, delicious! Okay, and what happened there? Well, we already have all that on tape, we have all invoices in the table and all it was thanks to this simplified automation that lacks invoices, which we no longer need agents, is a linear structure in which we only have we need a node of intelligence with very clear specifications. This can save you many manual labor hours by entering all these invoices. If you want like it flow, comment the word invoice and I will gladly send it to you.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250523 - Automatización telefónica básica para negocios que no necesitan grand... [7507604563092966698].eng-US_limpio.txt¶

There are businesses that with automation doing what the person did okay they don't need a big structure behind a pompous automation not the receptionist what they did was they took the call annotated on a sheet so-and-so such a date service plus nothing like this automation does the most basic of basics in automation let's try it out. Hi mute studio speaks isabella how can I help you today. Hi isabella how are you I want to schedule a hair removal for my wife it's a surprise. Hello, I am doing very well. Thank you. To schedule the appointment I need your name phone number and the service you want can you give me this information please. Yes my name is erick but her name isa. Thank you erick could you also provide me with the phone and what hair removal service you want fora she wants a brasilia and the phone number is five six one two two four four five perfect eric I will schedule her for brazil hair removal we will contact you to confirm what a nice surprise. Thank you for anything eric anything I am here to help you have a great day.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250721 - Chatwoot y n8n #automatizacion #zapier #ia [7529358057055259910].eng-US_limpio.txt¶

This recording agent has an advantage and that is that besides being able to receive images and voice notes it is connected with chatbot so all the messages arrive here in chatbot And best of all is that our agent has a tul which is called talking to a human When the agent detects that the customer is asking him to talk to a human he alone is deactivated and the agent changes from activated to deactivated and In this way, a label is added to our conversation called agent deactivated and here all the conversations where the agent was activated will appear and I, for example, if I have already solved the question and I want to return the customer to talk to the agent, I simply go here and put agent activated. messages are going back in through the agent and we can see how he is answering here in chatbot.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250808 - Solución automatizada para mejorar la gestión de cuentas por cobrar #... [7536091871731338510].eng-US_limpio.txt¶

This client had a problem, and that is that his department of accounts receivable was being a bit inefficient, not because they wanted to, but because the flow of invoices is already pending collection was too much. Then I propose this solution, which makes interday all invoices are distributed and all customers and all that information is looped. What does that loop do? What it does is that it processes each customer, check eh? Do you have outstanding invoices? And what it does is that with this if to say if you have, let's send an email using a template and using the data we have already collected from the customer. It is a simple template like this, which is a reminder and all these variables are replaced. and the email is sent. What does that do? Well, then we recorded that we sent that email, we unite here, we combine data together with customers who do not have accounts outstanding. And very important, we zoom in and out. The wait makes a wait at each iteration of the loop. If there are one hundred customers who owe, this is going to be a hundred times a loop ride to send check which ones should and in the end he will send a message of gratitude to the person in charge of sending that. With this, the customer saves hours of work. If you want me to share this template with you, let me have one of the primary colors below in the comments one and I will gladly give it to you

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250827 - He creado la solución técnicamente correcta para manejar los mensajes... [7543071507832917254].eng-US_limpio.txt¶

in a previous video we had talked about to solve the problem of when the user writes in parts using Redis and sleep was not the right way to do it. this time I come to talk about the solution the technically correct way to solve this is to use at least two flows one of the flows will contain all the automation logic while the other will include a Flood node. which acts as a gateway or gateway for all messages. this one would be responsible for grouping all the messages and also for avoiding overflow attempts that could cause the system to collapse imagine for example if a malicious user sends dozens of messages per second and let us bear in mind that the automations often work with AI agent nodes that consume tokens in this scenario, what would happen to the automations you currently have? if you want to see in detail how floodguard works how nodes are installed and how flows are structured I have the complete video on YouTube you can find the link in my profile don't forget to save this video and share it. as it will possibly be useful for more than one person to

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250812 - Hablamos sobre el curso N8N avanzado de cómo construir un chatbot más... [7537492396255350029].eng-US_limpio.txt¶

Hi I warmly welcome you to this new step in our learning I am very glad that you have decided to continue exploring the world of automation. If you are here it is because your profession has driven you to go beyond the basics and I am honored to be able to accompany you on this journey. In our previous course we built together a functional chatbot we laid the foundations but now we will build on that knowledge to create a much more powerful and professional one Throughout these lessons we will transform our bot into a real assistant with artificial intelligence First we will teach our bot to think he will learn a logic to handle much more complex conversations and multiple paths and decisions Then we will give him the ability to hear and see we will make him capable of doing so. We will build a bridge to the chatwood platform so that the bot can transfer the conversation to a real person when necessary. This is a skill that really makes the difference between a normal solution and a professional one. My goal is to guide you step by step in a clear and simple way. You don't need to be a programming expert, just bring your desire to learn and experiment. At the end of this course you will have in your hands an automated communication system that can generate immense value for any business. I am very happy that we start this journey together You are ready Let's go to the next lesson where we can see our starting point.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250824 - La clave para dominar n8n está en entender cómo funcionan los nodos： ... [7542173725098134806].eng-US_limpio.txt¶

If you understand this, you will have the basis to create any flow in n eight n because they all tend to share the same structure. A stream is always constructed with three types of nodes, trigger, processing and action triggers. The trigger is what triggers everything. For example, someone fill out a form on your website asking for a quote, that event ignites the flow. Then come the processing nodes. Here you can do things like cleaning up the data, use the guide to classify, search your crm if that email already exists, and many more that have to do with the preparation of the data we need to work. At the end are the action nodes, for example, save the id in notion, send you a notice by or send an automatic email to the customer. In short, it's like a factory. The trigger is when the raw material enters. Processing is the assembly line and the action is the finished product that comes out ready. If you understand this, you can create any flow you imagine, and in my community you have examples complete to copy and adapt. I'll wait for you there

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250831 - Chatbot de IA vendido que automatiza el 90% de la recepción de un hot... [7544725878761114902].eng-US_limpio.txt¶

I have sold this chatbot that automates ninety percent of a hotel When you order a room it tells you if there is availability it gives you prices answers frequently asked questions The most interesting thing is that you don't need humans at all. Even the ia retrains itself with all the previous conversations that right here is saved in a database. This is a real paradise. Follow me if you want to learn how to do this.

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250804 - Optimización de costos en N8N con APIs #N8N #automatización #chatbots... [7534571293908208910].eng-US_limpio.txt¶

Images of you are doing, but if what you want is to reduce costs in terms of ipis use, well, the it is advisable to first choose the model you are going to use. Remember that there are among the the models opening there are a lot of them, among them are the minis and the nanos. Nanos consume less than mini, but also must be seen new models consume less than previous models, then it depends whatever processing power you need, a chatbot with the nanos or the four point, is one nano enough to save you a lot of money? even eh? Optimizing the prom you use you can save some money and try not to use so many agents. Okay?

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250730 - Automatización simple para satisfacer las necesidades del cliente #N8... [7532943909396254007].eng-US_limpio.txt¶

Remember not all customers need extreme complexity This customer had the need for an artificial intelligence agent that answers all the questions of those of his customers who came from this facebook advertising that were always the same recurring questions So we installed this agent that does not have voice note support does not have image recognition support but works pretty well because it does what the customer needs. What does the customer need? Collect information from the people who write to him by whatsapp respond precisely to the customer's doubts regarding the business location services prices bases and contact modes The customer did not need forms did not need, did not need sales funnels Already with this automation he saves hours of his day That is why you have to study the customer's need before wanting to offer him asupermegahiperautomatizaci when he does not need it but of course we always ask for validations necessary for example those of multiple messages and multiple responses But then everything else is Remember that a person with a business with a venture does not need a dessert base does not need a cql base needs a table Needs an excel table where they can have all the information of their customers and that artificial intelligence people can also send you imes of some processes so not all the customers need complete automation. Is automating a part of the process sufficient?

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250730 - Hablamos sobre cómo la gente compra tiempo y soluciones personalizada... [7532681547082763575].eng-US_limpio.txt¶

Fully customized. And people that is consuming me, for example my service, it is because they do not want to create it themselves. Remember that people, people, buy time and there are people, for example he tells her, okay, a company owner of a company that sells plastics, selling metal sleeves, you come on the internet and say, give me a look, create a bot. Okay, bot builder create your own chatbot without code, but for him it is already complicated. Although there are days without snow, is complicated for him. You understand yes, yes, yes, yes, thousand. It scares me. Exactly this, for example for people who have experience. with code, who has experience with wordflow. Because remember that for this you have to have a little bit of logic, it is necessary to have a little bit of logic, of processes and not everyone has that. It's not, it's not common, is not the average, it is not the general. Then people look for the result. Look, I need a vote, you can do it. Yes, have my money, give me my bot and it works for my company. I win. You win. If you put a printer owner to do this that maybe is not texave, who is not a person who has taken on technology, is going to get complicated. The market that we are targeting is the type of person who does not have access, I provide you with a service, a finished product, continues

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250725 - Tips n8n #ia #automatizacion #n8n #zapier [7530823883662937350].eng-US_limpio.txt¶

This webflow here is a lifesaver. Why? Because it saves all your webflows in one Google Drive. I am very afraid of losing my webflows. in case something fails on my server, and what he does is that every week he creates a new folder on my Google Drive and then with the ene 8 ene API, because ene 8 ene also has an API and with this request he brings me all the waifus I have inside ene 8 ene. Then with the help of this code we convert it into binary files and already in that way we can upload them to Google Drive to the folder we just finished. to create, and it does so every week. So every week I can get involved to Google Drive and maybe search for a file that I am looking for or what such I lost or what I updated and now it fails. So I have a backup every week of my files. and the good thing is that those files are Jason code, so they weigh very little and they won't take up almost any space. If you want to learn how to do this workflow as well as other tips of n 8 n, comment tips in the comments and I will send you the video

/home/daniel/tron/programas/trans/vectorial/recopilacion_de_subs_limpios/bot/20250826 - Cuntio cuesta la IA？ #n8n #Productividad #inteligenciaartificial #age... [7541872178040540421].eng-US_limpio.txt¶

If you want to use artificial intelligence you must know how it charges. To avoid spending like crazy There are two concepts you should first know touch each other input and output taps go? The input words are all the words you write to GPT chat and the output words are all the words he replies to you. And companies usually have different costs for different models, For example for GPT 5 which is the new model the entry is aty.La $ departure is at $10 Here is the mini, here is the Nano gemina, the one from Google and this is Opus for example here, this one is very expensive, the entry is $15 and the exit is $75. And the cheapest is this one, the one with the GPT without the GPT Nano. Now let's make an example, for example analyze an email and from that email that extracts some data as name some features, It cost me 615 input and output tokens 1. 120 tokens. Now to understand these costs what you need to know is that these are measured per million tokens, that means that if this incoming email has 615 tokens to analyze it would reach 1. 626 emails. This is what it would cost us such a $1 25 analyze 1. 626 emails. Now let's analyze how much it costs to analyze with GPT 5, here are the operations in case you want to see them and here is the input cost which is point and $10 output. So this results in a point 0. 1 1. 9 by email dollars. This means that if we analyze 1. 000 emails the final cost will be $12 per and 1. 000 emails. But there is a way to make it even cheaper, What are we going to do, Let's switch to another model, to the GPT 5 mini, this the entry is $25 point and the exit is $2, that is quite, Here is the entrydayhere is point25 $2 $10. That is, if it changes we do the tun tun tun operations and we get a 0 point. 2 point 0. 0 $2 for each email analyzed. If we take it to 1. 000 emails will cost us $2, this means that it will be 80% savings, so before here it cost us almost $12 and here $2 for the same emails. By the same emails. So you have to keep this in mind when selecting which model you are going to use to automate or use artificial intelligence. If you liked this good content then follow here and I will try to keep explaining more artificial intelligence concepts.