I Built a Profitable AI Agent Crypto Quant - Here’s How (n8n)

here's how I built my own AI agent Quant for profitable crypto trading this is how it works you can type in the ticker of any major crypto so I'll do ethereum and wait for the workflow to execute our agent provides us with a detailed analysis on eth for either spot recommendations or for leverage recommendations if you want to be a bit more riskier so starting from the top under spot we can see that this is broken out between short-term and long-term so under short-term our Quant is recommending us to hold our eth and an entry price between 21. 85 to $21. 95 if we're looking to buy dips they also give a stoploss level of 2170 so that if we go below this level we should automatically sell and they also give takeprofit levels when we do make money and then all of these are based on three main rationals within the agent so the first is primary signals and these are based on just direct price action so we have things like trend lines support and resistance levels we can see that from the analysis the resistance is forming around 2220 and we have some indecisive price action as well then below this I also added in secondary indicators so these are the lagging indicators you typically see with macd uh RS I and obv so these aren't as important as direct price action but these also help in guiding the agent to give a more detailed analysis for us and then finally for the last rationale we have a sentiment analysis and this actually takes in all crypto news articles from the past 3 days and forms of sentiment on whether the short-term or long-term is bullish or bearish so right now we can see that based on news articles from the past 3 days the short-term sentiment is negative and then we have the same exact thing for the long term this is just on a longer time frame so now we can see that we have takeprofit levels of 2500 2700 and even up to 3,000 and below this we have leverage recommendations so they'll tell you positions to take as well as the amount of Leverage you should use so right now it's suggesting for us to open up a 3X long on eth with an entry price of around $21.

95 and a stop loss at 2170 now this is all for the shortterm we have the exact same positions but for the long term as you can see here I made this entire AI agent using n which is a no code low code Builder so I'll be going over in the next sections first how I built this at a high level so I'll explain all the nodes and then I'll actually build it from scratch so that you can customize it to your own liking and before we continue if you are interested in joining a community of AI Builders then make sure to check out our website at aiod pathways. com we help complete beginners build complex apps with AI as well as AI agents to use either for themselves or for their business so make sure to check out our website in the description if you do want to join a top AI building Community before going over the N workflow let me conceptually show you how this agent is actually calculating its analysis And Trades the analysis is made up of three components we have the primary analysis the lagging indicators and the sentiment analysis now going from the top to the bottom we actually start with live market data that we feed to the agent so these are hundreds of different Candlestick data and from there we can determine the primary analysis and the lagging indicators of all of this live market data then after this we grab a ton of different news articles that are all crypto and Market focused we run those news articles through a sentiment analysis node and then from there we can combine the primary analysis lagging indicators and the sment analysis to perform one final analysis now the reason why we don't just provide an image chart is because the Raw live market data can provide a more complete view into trading volume and prices over hundreds of different data points plus by providing the entire raw Market data the llm can determine its own signals and parts to the data on its own accord I personally found that image analysis of charts were a lot more flawed in terms of just giving an accurate response the reason why we don't just provide these standard technical analysis indicators like backd or RSI that most people reference is because they're all lagging and this means that they confirm price actions and patterns after they occur if you only wanted to use these technical analysis indicators you can literally just feed it to chat gbt you don't even need an agent for this but we want our Quant obviously to be more intelligent forward-looking and be able to actually predict Trends and this is where the raw Market data comes in now let's go over how I built this workflow at a high level we start with the trigger obviously that sets up our entire workflow in our case it's based on a telegram message but obviously you can change this to whatever you want if you're using slack or if you want to email or use any other third party application you can set the trigger up to be that specific node and then following this the next step in our workflow is converting whatever we type inside telegram into a specific format using a code node now this is because we're using the binance API in the next steps and it requires the ticker to first of all be in all caps and it also requires the trading pair attached so what this means is that if we type in BTC into telegram this code node will reformat it to say BTC usdt so BTC and all caps and then and usdt because that's the most common trading pair with Bitcoin and just to show you visually this is what the code looks like and if you're unfamiliar with coding that's completely okay I'll show you how to generate this code completely using AI when we build this from scratch it's honestly just as easy as dropping Anan nodes into your workflow so you can see we're first getting the input we're making it all uppercase we're adding usdt as the trading pair and then we're outputting this result so that whenever we type in BTC it'll output BTC usdt in all caps now once we have this cleaned up ticker we can send three separate HTTP requests to get Candlestick data for the ticker so for example if this is Bitcoin we'll get three different HTTP requests to get 15 minute 1 hour and one day candles for Bitcoin now the reason why we want to have three separate is because we want a comprehensive view of the coin we're looking at so the 15-minute gives us insight into 15-minute candles and prices for short-term analysis the 1 hour is more so for medium-term analysis and the one day is for long-term analysis and within each of these HTTP requests we're currently getting 200 items so if we were to double click here we can see that we want 200 separate 15-minute candles the same for 1 hour candles and one day candles so what this means is that every single time we run this workflow we'll get live candle data for the past 25 minutes for the past 200 hours and then for the past 200 days so this gives us a comprehensive view of how the live chart looks like from raw data and then the next step is just to get this Json output and label the data so right now if we click into the 15-minute candle request and then we test the step for Bitcoin all it does is give us a ton of data that's not labeled so what we want to do in the next step is take all 200 of these items and then attach 15 minutes to all of them so that we know these are all 15-minute candle data points so right now if we test the 15-minute candle data we can see that it adds 15-minute candles to every single one of these time frames so we know that these are all 15minute candles this makes it a lot easier to organize once we do this for every single different type of candle and merge them into one combined Json output so testing this data we can see that after we get the 15 minute 1 hour and one day candles it'll merge them into one Json output and inside this output we can see that they're all labeled with different time frames so these are all 15-minute candles if we scroll down a bit inside the Json file we can see that these are all 1 hour candles and then scrolling down a bit more we will see the one day candles as well right here so that's what this step is doing it's essentially labeling all the data for us before we merge it into one file and now like I mentioned this is all just one part of the analysis the agent also needs to factor in the current market sentiment and this is where we have our fourth HTTP request so after we receive all of this Candlestick data we'll then head over and then request some crypto news articles so if we double click into the request we can see that all we want is crypto or Bitcoin or coindesk articles and on the right hand side we can see that we have 2,000 different articles that we have to reference following this we just want to filter out the news articles just so there's less data we don't need things like authors uh your URLs image URLs when it was published or the content so all I have is title and description of the Articles this makes it easier in the next step to run a sentiment analysis and from the sentiment analysis we can see that they're taking all the Articles we published and it's writing out whether the short-term sentiment is either negative positive or neutral and a rationale for it as well as long-term sentiment and then now with this we can take all of our combined data merge them all into once again one Json output and as you can see this is what the final data looks like we have all of our candles with a different time frame so these are all our 15-minute candles if we scroll down a bit more we have all of our 1 hour candles and then if we scroll down a bit more we can see that we have all of our one-day candles here and then at the end of all of this is our sentiment analysis which once again was created inside the sentiment analysis node and this just gives us short-term sentiment whether it's negative positive or neutral and then the rationale and then the long-term sentiment once again negative positive neutral and the rationale and then it takes all of this inside One Singular Json output feeds it to our AI agent which then uses the Google Gemini chat model to produce our analysis and then I've just written out a custom prompt to spit out all the information we need so if I were to expand this we can see that we have the combined Market data with technical data and sentiment analysis saying that it's a Json array scrolling down we want first of all the agent to group The Candlestick objects and then we want them to analyze the data in detail with the short-term analysis long-term analysis and then trading recommendations below and then finally telegram has a 4,000 character limit so what I want to do is split the telegram message into two separate messages and then send them over to our telegram Channel now the first thing we'll want to do is add a trigger in our case we'll do telegram so we'll just simply type in Telegram and we'll want this whenever we send a message to the bot we'll do on message and I already have my telegram account set up but if you don't simply click docs and follow the instructions there they're pretty detailed and easy to follow along and then after this we can test the step to make sure it works so I'll open up Telegram and we'll type in test we can see that the test message did update with the text call test now like I mentioned next what we want to do is format our message to be something appropriate for the bance API so we can type in code and then add a code node to run some custom code okay so I just typed in BTC into our telegram trigger so we can see that what we want to convert is the text field called BTC you can see that the text field is nested within this message so message text this is exactly what we want to reference when we generate this new code now again you don't need to code at all here what you can do is simply ask chat gbt so I'll head over into chat jbt. com and then inside chat gbt all I did was type in please WR JavaScript code for my nan code node that takes an input remember again it's nested within message so it's message.

text if we look here it's called text the field but it's nested within message that's why we're doing json. message. text make sure it trims the white space and then converts it to all upper case and then what we want to do is add usdt to form the complete trading pair symbol so like I mentioned this is the proper format for binance to recognize our HTTP request and then we just want to Output the results as Json so we can see to create an NA and code node code they'll provide us with the code and it's as simple as just copying over this and again we don't even need a reasoning model I'm on chat gbt 4.

5 but you can do something like 40 this is a fairly simple request if we head back into our workflow copy over the code test the step this should take the text BTC and then first make it all uppercase remove whitespace and then add usdt to it and if you do want to know what's going on in the code they've added explanations above every single line so it first retrieves the input text so this is Json message text which is this BTC it first checks if it's a string this isn't really needed but they did it uh next they want to trim the white space and convert it to uppercase so what we want to do is trim convert to uppercase and then adds the word usdt to form the trading pair symbol so it takes our trim text and just adds usdt and then it outputs the result as a Json object and this is what we see here which is just BTC usdt now next what we want to do is an HTTP request to get Candlestick data so we can just add one here HTTP request and we can simply just use the binance API because it's free so the method is get because we want to get data now you can find this exact URL inside the documentation or ask AI but I've just copied and pasted it here to be a bit more simple so it's api. binance. com API vtk lines and then for the symbol we want this to be the simple we type in so in our case it's BTC usdt and the interval like I mentioned we want three different intervals we want 15 minutes 1 hour one day so the first is 15 minutes and we want the limit of 200 so this is a set of a limit of 200 we can set this to be 100 all the way up to to 1,000 so right now if we test this step it should take BTC usdt and provide us with 200 different items and this is formatted all inside the Candlestick data so if you want to look at this this is how it reads we have the open time open high low close and volume so this is what these all mean now next what we want to do is actually remove this BTC because we want this to be dependent on what we typed in the telegram message before so what we can do is make this variable instead of hardcoded so rather than having this just read BTC usdt every single time let's delete this and then make this into an expression so if we exit out of here let's click expression go to our schema and then literally drag in the trading pair symbol this way the symbol will always change based on what we type in in the previous node so right now these are all the Open high low close prices for Bitcoin but let's test this out with a different type of crypto so heading back to our telegram trigger let's first test this and then I'll type in eth press enter we can see it's executing so inside the code node it first changes eth to eth usdt and then for the HTTP request this should now all the eth prices perfect because now you can see it's outputting eth and this symbol is changed to eth usdt now like I mentioned these are just the 15minute candles so let's rename this so we know what this is this will be HTTP 15 minutes and now let's copy and paste this two more times for the 1 hour and one day candles so now that we have the 1 hour one day let's connect this code node to both of these and then what we want to do is adjust this URL so rather than having this be 15minute intervals we now want to set this to be 1H hour intervals and once once again we can change the limit if we want so if you want to pass over the past 1,000 hours we would change this 200 to 1,000 and then finally in the one day just double click and then change the interval to 1 day instead of 15 minutes so now testing this out we can see that we have 15-minute candles 1 hour candles and one day candles and they all output 200 items so these are the milliseconds and each millisecond is 1 hour apart for the 1 hour candles and these are all one day apart nowhere inside this Json output does it say one day candles anywhere so that's what we need to do in The Next Step label all the data so next what we want to do is just simply add a node for edit Fields cuz we want to edit all this data all 200 items so type in edit we can do edit Fields set and we're working with Json so we can just set the mode to be Json and then all we want to do is just append 15 minutes to each of these data points so what we can do is set a Json expression all we're doing is doing time frame 15 minutes and then adding all of these candles so you can copy this exactly and we can see that every single candle now says 15 minutes and they do this for all 200 items so we should have 15 minutes labeled for every single item 200 times next let's exit out and do the same exact thing for the 1 hour in one day so for the 1 hour set this to be Json once again we'll set this to be an expression and we just want to add in the 1 hour time frame we'll do 1 hour test the step and we can see that every single candle now has 1 hour attached to it so we know that these are all 1 hour candles and then once again for one day again we'll edit Fields set the mode to Json delete this set this as an expression and we want this to be one day candles if we test the step every single one should now say one day and again this is just labeling the data so that the AI agent once we feed all this data knows that these are 1 hour one day or 15minute candles and then next what we want to do is aggregate all this data to make it easier for the AI agent to receive and understand so we can combine all of these 200 items into one single Json output so what we do is add a merge node so just type in merge and then make sure the mode is set to aend we don't want to combine and merge any matching items together we want this to be one after the other so it'll first show all the 15-minute candles and then all the 1 hour candles and then all the one day candles so select the pend and because we have three inputs we'll select the number of inputs 2 B3 and then next all we do is connect the inputs so first input to the first field second to second and third to third so now let's test our workflow from the start we'll do test and then I'll just type in eth once again this should trigger this should send a 15- minute 1 hour one day candles 200 items for each so now this merge has 600 items if we so because we did append it starts off with all the 15minute candles if we scroll down somewhere to like at the middle we have all the 1 hour candles and then scrolling all the way down we have the one-day candles next what we want to do is combine all of these 600 items into one single Json output that makes it easier for the AI agent to kind of digest and understand so once again we'll add another code node so from here just type in code and then we can run custom JavaScript or python code so inside this code node once again we can go to Chad jbt I've typed in the prompt I'm working in an N workflow and have 600 candle items from a merge node so that means is we have 600 items from this merge node and then what we want to do is combine them into one single output so I'll say each items Json already represents a single candle I want you to combine them into one output item with a property called all candles it's an array containing all 600 candles once again so simple all you want to do is just copy over the code paste it into your code node here and then test the step and it should take this Json and then combine it all into one items right now it's 600 items each item representing their own candle and then now we have One Singular Json output but as you can see all the data Remains the Same we have all our 15-minute candles here scrolling down 1 hour candles and then scrolling down we have the one day candles now just taking a step back this is our current workflow we have a telegram trigger code and then we have three different HTTP requests so 600 items total we're merging those into One Singular Json output and then combining those 600 items into One Singular item so let's go ahead and just rename this code node to be combine Json data now that we have all our Candlestick data done what we want to do is add in a fourth HTTP request for our sentiment analysis so like I mentioned we want to get relevant news articles and then run it through an AI node so just head over to newsapi.

org and you can get your own API key after you sign up for an account then now heading back into our workflow let's go ahead and add in our fourth HTTP request so once again just type in H HTTP request perfect in the method we want to get data once again and for our URL let's go ahead and just type in news api. org fre2 everything so again this is the URL to call on the news API and once again you have to make an account and get your own API key but from here what you want to do is first send query parameters and these are the parameters in which we can search for articles and also include our API key so the first parameter we want to include is titles so we'll call this q and we want this to be all news articles or titles that contains the word crypto or Bitcoin or coindesk the reason for this is because we want all crypto related articles Bitcoin related articles and then coindesk is one of the largest crypto news networks so you want all articles by them as well now this news API Aggregates all news articles so if you were to just run this query parameter it would literally come up with millions of different crypto articles so what we want to do is add a time parameter so that it's only recent articles so what we can do is call the next one from and what we want is all articles within the past three days now obviously you can change this to whatever you want but in my case I'll just set in an expression for today's date minus 3 days so all you can do is go ahead and copy and paste this function that I wrote but it's basically the date now minus 3 days so right now today's March 7th 2025 this would be March 4th 2025 and what this means is that it's taking all of the news articles about crypto or Bitcoin or coindesk from March 4th 2025 up till today which is March 7th 2025 next we can sort these by popular so once again add another parameter and what this does is bring the most credible articles to the front so that they're factored more heavily inside the sentiment analysis so we call this next one sort by and then for the value we'll just call this property or [Music] popularity and then lastly what we want to do is insert our API key so type in API key and then type in your own API Keys value so we want to make sure that this HTTP request is wired up so what we can do is just connect this to the code node so that it's going over all four of our http request now the way this workflow works is that it'll first execute these before it moves on to this fourth HTTP request so I'll show you what this looks like after we run this once again but for now let's go ahead and test this to make sure that this is working so if we were to test the step great so we have 1,265 articles and they all have to deal with crypto so you can see the first one Trump agencies drop dozens of Biden a cases against crypto see Trump admin Trump tariffs trade War this is all Market crypto and Bitcoin related now next what we want to do is filter out this data you can see we have things like author uh url url image publish that these are all points of data that we do not need and if we have 1,000 results of this this is literally just increasing our output to the AI agent for no reason now obviously this is important just in terms of context so what we can do is just delete essentially every single field except for the title and description since these contain all of the content about what the article is about so once again we can add a code node so exit out of here we'll add one item we'll call this code and then what this will do is essentially filter out the news articles so we can rename this to be filter news articles now again these aren't complex lines of JavaScript code so once again you can head over to chat jbt or your AI chat bot of choice ask them that you want to filter out your articles to only include certain Fields so in our case here I'll just paste this in we have a new one called filtered articles and it only includes title and description so right here we can see that we have author title description url url image publish that content then now if we test the step you can see it only returns title and description make sure it works filter articles title description title description title description so you can feel free to copy my code or if you want to generate custom one just simply ask AI again these aren't complex lines of code so pretty much everything will work now next what we want to do is actually run a sentiment analysis based on all these filtered news articles now the reason why we want to run this analysis separate from our further analysis later on with an AI agent node is because we want our data to kind of be broken down so we want essential AI agent node to focus only on a sentiment analysis on articles and the next one to take the output of what we received here which is the sentiment analysis so that it has an easier time to essentially combine all this Candlestick data with an already finished sentiment analysis so let me show you what I mean all we can do here is add in an AI agent node now I'll just add in open AI since I like using open AI it's generally for things like this but if you want to use a different one you can select basic llm chain but in our case we'll just do message model for open AI now just connect to your account if you haven't done so already resource will be text message a model and because this isn't too complex we're just taking literally hundreds of Articles and asking them to run a sentiment analysis on all of these we don't need a complex model we can just do 40 mini and then what we want to do is add in a system message to set the model's Behavior we also want to Output the content as Json and then for the text once again you can have ai generate things for you but in my case I'll just type this in and feel free to copy it just says you're highly intelligent and accurate sentiment an analyzer we have short-term sentiment and long-term sentiment so I'll pause it here and then stop here as well for the long-term sentiment section and then this just gives an example of what the output should look like so we have a short-term sentiment category score rationale long-term sentiment category score and rationale now let's go ahead and test the step so we're inputting all of these different news articles and the output should now be two different sentiment analyses for short-term and long-term great so we have short-term sentiment negative . 6 we have the rationale and then long-term sentiment is neutral and we can see that the rationale is including all the articles that we fed it so we can see that key events include Trump's proposal for a strategic crypto Reserve we also see fears of rising tariffs and a potential trade War which once again are all recent articles that just occurred and the long-term sentiments once again has the same exact information but on a longer scale so now looking at this in aggregate what we have is all the Candlestick data merged into one single Json output and then what we have is all the filtered news articles and what we're doing is running a sentiment analysis through an open AI message model and receiving one item as well then just so we're organized let me just rename this to be sentiment analysis and then just as a quick aside let me delete this line what you don't want to do is just use the built-in sentiment analysis that n gives you because this isn't that detailed and it has a very strict system prompt so that you won't be able to get actual text explanations as to what is the rationale behind the figures so this node will give you scores as well as categories whether it's positive negative or neutral but they can't actually display text for you in any meaning way that's why I just like to bypass this entire sentiment analysis node and just make our own through a message model in our case open AI now we should have one item from here and one item from here but let me just run the workflow once again just so you see that this is one item at the end and one item here okay so now that we both have one item what we need to do is merge the data so it'll be merged into two items and then once again combine that into one single item so once again add in the same exact merge node so what we want to do is append this time with two inputs because we want to append the combined Json data which is all the Candlestick data with our newly created sentiment analysis so with this there's two inputs in the merge to now create two separate items now let's go ahead and test the step and we can see that it creates a Json output with two items the first item is all the candles starting from 15 minutes if we scroll down we have the one hour candles up here and then we scroll down a bit more we have the one day candles shown here and then if we scroll down all the way at the end it should have our sentiment analysis so again we're not passing all of the news articles to our AI agent at the end all we want to do is pass over the sentiment analysis that was already calculated just so we don't need to do like 20 different calculations all at once so with this the AI agent essentially doesn't calculate anything they just read what the rationale is and they know whether the short-term sentiment is negative or positive and whether the long-term sentiment is negative positive or neutral so exiting out of our merge node once again we only want to send one item to the AI agent at the end so what we can do is once again combine all the Json data so add in a code node it might seem redundant but it is necessary so go ahead and rename this to combine Json data once again Ask chat gbt but it's a pretty simple JavaScript code all we want to do is initialize all candles content data scroll down and then it return the single item with both Candlestick data and content now let's go ahead and test the step and as you can see from the merge we have two items and now we have one single item which once again starts with all the candles so we have 15minute candles scrolling down just to check once again we have 1 hour candles in the middle and then scrolling down we have one day candles near the end so 25 minute candles 200 1 hour 200 1 day and then at the end of all of this is our senent analysis now this is nicely formatted so that we can finally send this one single item to our AI agent node so let's now click add another node Advance Ai and we'll do AI agent now this AI agent node here is where all the magic happens we can finally take all the data we aggregated from this Json output and then actually get something meaningful out of the agent all right so the first thing we need for the agent is a brain and because this is a ton of data that's coming in we should choose a more powerful model so if you do have credits to spare I would maybe do something like chat gbt o1 or 03 but I like using Gemini because it's free so selecting Gemini here what you want to do is first sign up for your account and then I usually do 2.

0 Flash and then next what we need to do is carefully craft The Prompt for this AI agent so that they understand all the information that's coming in and because we're having the agent just analyze all the data we don't actually need any tools we can just select conversational agent since they're talking back and forth to us next for the source for prompt we want to enter in our own prompt so we'll select the find below now this is the custom prompt that I made but feel free to copy this prompt exactly or tailor it to your own liking but going over this what I first do is say here's the combined Market data for and then whatever we typed in for the trading pair symbol so this is referencing our Bitcoin usdt all I did is drag this over so that we now have Bitcoin usdt here on the right hand side and this right hand side is what the agent sees so right now we're referencing some data points inside our previous nodes so right now it sees that here's combined Market data 4 and then because we typed in Bitcoin in the trigger it'll see Bitcoin usdt and then below this I provided all the technical data and sentiment analysis so if you recall earlier when we combined the Json data we have two sections all the Candlestick candles as well as all the content from the sentiment analysis so what I did was drag this over drag this over and we can see that on the right hand side for the output this is what the agent sees they see all the technical data and then below all of this we have the sentiment analysis so this shows short-term as well as long-term sentiment so this is just make sure that the agent is actually referencing the correct data points inside our previous nodes and then below this I'm just typing in some general context for the agent this is so they know how to read the time frame Candlestick data and this is so they know how to read the sentiment data as well and then below this we have the actual instructions for the agent so we want them to first group The Candlestick data into either short-term medium-term and long-term then we want them to analyze the data in detail for short-term and then do the same for long-term analysis and then after they've computed all the analysis in the back end we want them to generate trading recommendations for both spot Trading which we saw in the beginning of the video as well as for leverage trading below so obviously we have things like position leverage entry price stop- loss and take profit levels finally below all this we want the output format so we want to make sure this is formatted properly for telegram so I'm adding things like Bolding uh line breaks and then also adding in the date so Bitcoin analysis for today's date and then below all of this I provided an actual example of what the output should look like and the same for leverage recommendations so now exiting out of this massive prompt let's go ahead and test this step and awesome we have the output now so as the Bitcoin analysis for March 7th 2025 and we have all this markdown formatting because again we want to make sure it's fine for telegram so this might be a little bit hard to read inside this output window but as you can see we have spot recommendations below this we have all of the different types of actions entries stop loss take profit and rationals we have the same for longterm as well as leverage trading below now all we need to do is take the output from this agent split it into two different messages because again this is way too long to send in one single telegram message so to split the message into two super simple what we want to do is just add a code node and then you can either copy and paste my format exactly or you can just have chat jbt spin something up for you again this is super simple JavaScript code and now it splits it into two different messages this way we can output both of this content into telegram so now finally as our last step let's go ahead and send this all to telegram so we'll select Telegram and then we'll scroll down and select send a text message we'll keep these three all the same and for the chat ID we want to reference the initial chat ID all the way at the beginning of our workflow so what we can do is double type into this scroll all the way down to our initial telegram trigger and then what we want is this chat ID so just simply copy and paste it in and then the last thing you need to do is rather than item you have to type in first so this just Returns the first output from an index and you get the result which is 7596 which is the same exact chat ID and then next for the text what we want to do is directly output the contents from our last node and then lastly what you want to do is make sure that this parse mode is set to HTML this just copies over all the form formatting that we have from the previous steps so that we bold and have proper line breaks and then what we can do now is test the step and awesome so now it's sent over to items to telegram so we have the first half of the message here we have what appears to be all of our formatting and then the second half of the message below it here so just to show you that this works this is Bitcoin analysis for March 7th at 10:17 p. m. we have the same exact thing here 10:17 p.

m. we have the spot recommendations shortterm all the rationals long-term all the rationals leverage recommendations as as well for shortterm and longterm and we can see how it's split into two messages so that we're able to pass over all this information to telegram now like I mentioned if you do want to join a top AI building Community where we take beginners and help them build either complex apps with AI or AI agents for their business then make sure to check out aiod pathways.