{"id":"2093378281810833789","url":"https://x.com/everestchris6/status/2093378281810833789","text":"https://x.com/i/article/2093378275301273600","author":{"name":"Chris","username":"everestchris6","avatarUrl":"https://pbs.twimg.com/profile_images/1970372109202530305/A5Axp4ih_200x200.jpg"},"createdAt":"Fri Aug 28 16:40:50 +0000 2026","engagement":{"replies":3,"retweets":7,"likes":66,"views":41104},"article":{"title":"make money with grok bot in 6 weeks (FULL GUIDE)","previewText":"you've got 6 weeks before this window shuts. here are five systems you can build with grok bot that'll pay you through christmas.\nthey sell into one trade with jobs up to $8,000, and their whole","coverImageUrl":"https://pbs.twimg.com/media/HQ0sbE-awAARfTX.jpg","content":"you've got 6 weeks before this window shuts. here are five systems you can build with grok bot that'll pay you through christmas.\n\nthey sell into one trade with jobs up to $8,000, and their whole season books out in the next six weeks. you get paid monthly through it, so what you set up in september is what pays out in december.\n\nby the end of this guide you'll have:\n\n- five separate systems you can sell into one trade\n\n- a render pipeline that puts lights on a real house\n\n- postcards printed and mailed autonomously\n\n- a way to price each system so it actually pays\n\n- the installer list built out of public data\n\n- grok bot running all of it on a schedule\n\ninstallers sell two things. putting lights up for the season, which averages about $443 a house, and permanent lighting that stays on the house all year and changes colour, which runs $2,000 to $8,000.\n\n![](https://pbs.twimg.com/media/HQ0sbT1bcAAGGMI.jpg)\n\nwhy postcards and not ads\n\nthis trade is bad for paid ads and good for physical mail.\n\nthe buying decision is made by a homeowner looking at their own roof, and the thing that sells it is seeing their own house with lights on it. you can't customize that in a feed. you can do it on a card that arrives with their address on it and their house on the front.\n\nevery installer in a metro is bidding on the same keywords. nobody is really doing physical mail.\n\nand the data is free. street view and satellite imagery cover every house in the country, and public address data tells you who lives on each street. that's the whole input for this.\n\nstep 1, the setup\n\ngrok bot has its own computer with a terminal, it holds the print account, the payment account and the render pipeline itself, and none of it needs you after the first run. use a print and mail provider with an api, so grok bot can send the cards itself.\n\npaste this:\n\n> set up everything you'll need to run postcard campaigns for me later.\n\n> \n\n> - here's the print api key: [key]. read their docs and tell me the exact card size, bleed and dpi they want, and generate everything to that spec from now on\n\n> - set the monthly spend cap at the print provider's end, not in this chat\n\n> - here's the payment account. create products and checkout links from the terminal, never in a browser\n\n> - set up address validation before anything prints. a card to a bad address costs the same as a good one\n\n> - set up a google cloud project and enable the street view static api and the maps static api, then give me one key restricted to just those two\n\n> - tell me what a batch of a thousand addresses costs at four street view frames and one satellite view each, after the free monthly allowance\n\n> - before you print anything for real, send one test card to my own address and show me what came back\n\n> - tell me now what you can't do with the access i've given you, rather than working around it later\n\nstep 2, the render\n\nyou need to take an address, get an image of that house, and put lights on the roofline in a way that a homeowner looks at and recognises as theirs.\n\nthe images come from two google apis. the street view static api takes an address and returns a photo of it, and you get a different angle each time by changing the heading value in the url. \n\nthe maps static api returns the satellite view of the same address with maptype set to satellite. both are simple url calls, and grok bot can call them straight from its terminal.\n\nthere's a third endpoint that matters more than either of those. street view has a metadata endpoint that tells you whether imagery exists for an address and what date it was shot, and those calls are free and don't count against your quota. \n\nso you check every address for free first, and only pay for the ones that come back with something recent enough to use.\n\nthe reason you pull several frames rather than one is that a single photo leaves the model guessing at everything outside it. four or five images of the same property and it has enough to work from.\n\npaste this:\n\n> build me the render pipeline. it takes an address and returns an image of that house with christmas lights on it.\n\n> \n\n> - check the street view metadata endpoint first for every address. those calls are free, so use them to throw out anything with no imagery or imagery older than about three years before we spend anything\n\n> - for the addresses that pass, pull several street view frames from the static api at different heading values so we get the house from more than one angle\n\n> - pull the satellite view of the same address from the maps static api with maptype set to satellite\n\n> - pass all of them to the image model as reference so it has enough of the property to work from\n\n> - use an image edit model (preferably chatgpt image-2) so the house stays exactly as it is and only the lights are added. do not regenerate the house\n\n> - the output should look like a photo of that house taken at dusk with the lights on, not an illustration\n\n> - lights follow the roofline, the porch and the garage edge. warm white by default\n\n> - do not add lights to trees, bushes or anything the installer might not quote for\n\n> - return the address, the reference images and the render together so i can check them side by side\n\n> - tell me what the whole batch cost in api calls when you're done\n\n![](https://pbs.twimg.com/media/HQ0sbk8bEAAbfj0.jpg)\n\nrun twenty addresses through this and look at every one before you go further. just to check if everything looks good.\n\none line to keep in mind through all of this. a render is a visualisation and not an actual quote. it shows what the house could look like, and the installer prices the job. don't build a card that implies otherwise, because the installer is the one who has to deal with it.\n\nstep 3, finding the installers\n\nyou're selling to the installer, not the homeowner, so you need the list first.\n\npaste this:\n\n> build me a list of christmas light installers in [region].\n\n> \n\n> - search google maps for christmas light installation, holiday lighting, and permanent lighting in every city over 50,000 people in the region\n\n> - also pull landscaping and holiday decor companies that list light installation as a service, they're most of the market\n\n> - for each one: business name, owner name where public, phone, email, website, review count and rating\n\n> - flag whether they offer permanent lighting or seasonal only\n\n> - flag whether their website mentions taking bookings online\n\n> - drop anything with under 20 reviews or over 200, and anything that's a franchise\n\n> - tell me which of these look like one man and a ladder rather than a business with crews\n\nsomeone doing twenty houses a season wants leads and will never pay a retainer. but on the other side, someone running a couple crews knows what a lead costs him already and will.\n\nthe permanent lighting flag is the other one. an installer already selling permanent is a different customer from a seasonal-only one, and system two only works on the first kind.\n\nthe five systems\n\neach of these is a separate product with its own buyer and its own price.\n\n![](https://pbs.twimg.com/media/HQ0sb1DaMAYxFyu.jpg)\n\nsystem one, the street campaign\n\nyou pick the streets, render every house, mail the cards, and the calls go to a number that's yours.\n\nstreets are chosen on evidence. you want a neighbourhood where people already spend on this, and there are three public places that tell you which ones those are. every city has a crowdsourced christmas light map that people update each december with addresses on it. local news runs a best streets roundup every year. and the same question gets asked in local groups and on reddit every year.\n\na street that shows up in more than one of those already decorates, which means it already spends.\n\npaste this:\n\n> find me the streets worth mailing in [city], then build the campaign.\n\n> \n\n> - pull any crowdsourced christmas light map for this area from the last two years, plus local news roundups of the best streets, plus what comes up in local groups and reddit\n\n> - give me the streets that appear in more than one source\n\n> - ask the installer which streets they already know before you use any of this, theirs will be better than anything you find\n\n> - for the streets i pick, pull every residential address\n\n> - drop apartments, anything with no clear front elevation, and any address that failed the render check\n\n> - render each house, then generate a card with that render on the front\n\n> - the back has one line about lighting on that street this year, the installer's business name, and a phone number\n\n> - leave the qr corner blank and tell me, i'm compositing a real one in\n\n> - prepare the batch fully and stop. i approve before anything goes to print\n\n> - tell me the cost of the batch before you print it, not after\n\nhow you charge for this is your decision. a % of each closed job sounds right but it kinda isn't, because you can't see what closed unless you build a proper tracking system.\n\na retainer for exclusivity in a set of postcodes is cleaner. the installer is the only one you mail for in that area, they pay monthly through the season.\n\nsystem two, the permanent lighting version\n\nsame system, different houses and a much bigger number.\n\npermanent runs $2,000 to $8,000 against a lower amount for seasonal. one closed permanent job is worth eight seasonal ones, which changes what you can afford to spend to get it and changes who you mail.\n\npaste this:\n\n> run the same campaign for permanent lighting instead.\n\n> \n\n> - go through the reference images and pick out the bigger houses with a long straight front, since those are the jobs worth having at this price\n\n> - prefer newer builds and streets where houses are clearly owner occupied rather than rented\n\n> - the render shows the lights in a colour other than warm white, so it reads as year round rather than christmas\n\n> - the card leads on it staying up all year and never needing to be hung again\n\n> - only mail this for installers who actually offer permanent. check their site and ask them before the first batch\n\n> - work out the cost per card and tell me how many closes it takes to pay for itself at their average job value\n\nthe colour change on the render is the pitch. the point of permanent lighting is that it isn't only for christmas.\n\npermanent is priced on how much of the house gets covered, so a big house with a long front is a bigger job than a small one.\n\nthis is also the one that keeps working after december, which matters because the rest of the systems go quiet after december.\n\n![](https://pbs.twimg.com/media/HQ0scCGaMAAkKUv.jpg)\n\nsystem three, last year's customers\n\nthis is the easiest sale in this trade.\n\nevery installer has a list of addresses they lit last year. those people paid, the install went fine, and most of them have not been contacted.\n\npaste this:\n\n> the installer has given me their customer list from last season.\n\n> \n\n> - render every address on it the same way, using their real install photos as reference where they exist\n\n> - two versions of the card. one to rebook the same package, one offering permanent instead\n\n> - send the permanent version to anyone with a bigger house who paid above their average last year\n\n> - do not include any pricing on the card. the installer quotes\n\n> - schedule this to land in the first week of september, before they start booking elsewhere\n\n> - flag any address that appears more than once, they've had multiple jobs and go in their own batch\n\nyou charge a setup fee plus a monthly here, and it's an easy conversation because you're using their own list and their own customers. there's no targeting for them to doubt.\n\ntheir own install photos are the best reference you'll ever get for these renders.\n\nsomeone who has booked twice is the person most likely to buy permanent, and they should get a call rather than a card.\n\nsystem four, the quotes that never closed\n\nevery installer has a pile of people who asked for a price and never responded.\n\nmost installers have this this data, so the first job is finding it.\n\n> the installer says they don't have a list of quotes that didn't close. help me find out whether that's true.\n\n> \n\n> - ask them which app they used to send quotes last season, and whether it exports\n\n> - ask whether they use a crm, a scheduling tool, or an invoicing tool, since all three usually hold the quote history\n\n> - if they quote by text or email, ask whether they'd give me a search of those for the months they were quoting\n\n> - if the quoting tool has an api or a csv export, tell me exactly which fields we can get and which we can't\n\n> - the only fields that matter are address, date, and amount quoted. everything else is optional\n\n> - if there is genuinely no address on any of it, tell me that plainly rather than trying to reconstruct addresses from names\n\nonce you have the list, the campaign itself.\n\n> take the installer's list of quotes that didn't convert.\n\n> \n\n> - render each address the same way\n\n> - the card references that they looked into this last year, without saying what they were quoted\n\n> - keep the message short. the render is doing the work\n\n> - separate anyone quoted above the installer's average into their own batch, they get the permanent offer instead\n\n> - drop anyone quoted more than two seasons ago, and anyone whose address also appears in the customer list, they booked in the end\n\n> - mail this in late september, after the rebook batch has gone\n\n> - tell me how many of the addresses failed the render check before you print, not after\n\nthe reason this converts better than a cold street is that the objection last year was a price against an imagined result.\n\nprice this one lower than the rebook campaign, because the list is smaller and colder. it works better as an add-on to system three than as its own sale.\n\nsystem five, the off-season list\n\nthis is the one you build all season without doing any extra work, and sell in january.\n\nevery enquiry that comes into your campaigns and doesn't turn into a job is a person who wanted lights and didn't get them. basically, some of them called too late, some couldn't get a slot, some got quoted and went quiet. all of them are still homeowners who wanted this.\n\nwhat makes the list worth money is that it can't be scraped. anyone can buy a list of addresses in a postcode. nobody can buy a list of people who wanted christmas lights in november and didn't get them, because that only exists if you were running the campaigns.\n\n> keep a running list of every enquiry that came in and didn't turn into a job.\n\n> \n\n> - date, address, which campaign it came from, what they asked about, and how they got in touch\n\n> - tag whether they asked about seasonal or permanent\n\n> - tag the reason it died where i've told you: too late, no slot, price, no reply\n\n> - keep it separate per region so i can sell it by area\n\n> - never include anyone who asked not to be contacted, and keep that suppression list forever\n\n> - do not merge this with any list an installer gave me. theirs is theirs\n\nthen in january:\n\n> it's january, get the lists ready to sell.\n\n> \n\n> - for each region tell me how many are on it, the split between seasonal and permanent, and how fresh they are\n\n> - work out what the list is worth at that installer's average job value and a conservative close rate, and show me the working\n\n> - do not present a close rate as a promise. show it as the maths and label the assumption\n\n> - build a one page summary per region with the counts and the split, no addresses on it\n\n> - tell me which regions have too few to be worth selling rather than padding them\n\njanuary is the only month these people will talk to you. they're planning next season. \n\na list of people who wanted lights a few weeks ago and didn't get them is the easiest thing anyone will offer them all year.\n\nyou can sell it as a flat price per region.\n\nthe one line to keep honest is the close rate. you don't know what these convert at until an installer has run them.\n\nhow you price all of this\n\nthe retainer is the right shape for most of it. the reason is that you cannot see which jobs closed. the call goes to a number, the installer drives out, quotes on a driveway, and either closes the deal or doesn't. \n\nso charge monthly and take exclusivity off the table as the thing they're buying. you mail for one installer per set of postcodes. \n\nthat's a real asset, and it's the reason they will hire you again next year.\n\nthe number itself comes from what a lead costs them today, not from what an agency charges. contractors on lead platforms have written about paying $188 for eight leads and getting no jobs.\n\nand another about $520 for two. that's what's in their head when you say a price. if your monthly is less than what they're already burning on a platform that sells the same lead to four of their competitors, it's super easy to close them.\n\nthe maths you should be able to do out loud on the first call. a seasonal job averages around $443. if a batch of a few hundred cards produces even a handful of jobs, the retainer has paid for itself and the rest of the season is profit for them. \n\nyou don't need to promise a conversion rate to make that point, you just need to show them the number their own average job value implies.\n\npermanent lighting is where you can charge more for the retainer. those jobs run $2,000 to $8,000. one close pays for a whole season of what you charge, so price the system accordingly.\n\nfor the last-year list in system three, charge a setup fee plus the monthly. the setup covers the render pass over their whole customer list, which is real work you do once, and the monthly covers the batches going out through the season.\n\nsystem four can be an add-on rather than a standalone. the list is smaller and colder, so it goes on top of system three at a lower number rather than being sold on its own.\n\nsystem five is kinda the odd one out because you're selling an asset, not a service. so a flat price is better for this.\n\ntwo things to avoid. don't do a free trial batch, because printing and postage is real money.\n\n![](https://pbs.twimg.com/media/HQ0scTpaMAM742T.jpg)\n\nthe first approach\n\nshow up with their own work already done. before you contact anyone, render four or five houses in the area they operate in and have those ready.\n\n> get me ready to approach [installer name].\n\n> \n\n> - pull their site, their google listing and their social, and tell me what they actually offer and what they charge if it's public\n\n> - tell me whether they offer permanent lighting, and whether they're pushing it or just listing it\n\n> - read their reviews and tell me what customers say goes wrong, in the customers' words\n\n> - pick five streets in their service area where houses look like they'd spend on this, and render one house on each\n\n> - write me three lines i could open with, using what i learned rather than anything generic\n\n> - tell me anything about this business that suggests they can't afford a retainer, rather than letting me find out on the call\n\nyou send four images of houses in their area and ask whether they've ever mailed to that street. you're not describing anything, they're looking at it.\n\n![](https://pbs.twimg.com/media/HQ0scl3a0AAceRr.jpg)\n\nthe dashboard\n\nbuild this to track all the data.\n\npaste this:\n\n> build me a dashboard and host it. give me a live url\n\n> \n\n> - one row per campaign: which installer, which region, how many cards, when they landed, cost\n\n> - a second view for the render queue, showing every address with its reference images and the render side by side and a pass or fail button\n\n> - a third view for enquiries, with which campaign they came from\n\n> - keyboard shortcuts to move through the render queue and pass or fail\n\n> - tell me now if the host you're planning to use can't keep this url alive permanently\n\nevery batch needs a human looking at the images before they print.\n\nhaving the reference images next to the rende makes that quick. you're basically just checking if it's the same house.\n\nwhat it's allowed to do without asking\n\nset this before you turn on the schedule.\n\npaste this:\n\n> these are your permissions and they don't change without me saying so.\n\n> \n\n> - you can do these on your own: pull addresses, render, generate cards, build lists, draft copy\n\n> - you prepare and stop for anything that prints, mails, or spends money\n\n> - never print a batch i haven't approved, and never approve one yourself because a deadline is close\n\n> - never contact an installer or a homeowner directly. i do that\n\n> - never post from any of my social accounts\n\n> - if a render fails the address check, hold the whole batch and tell me rather than dropping that one card and printing the rest\n\nthe schedule\n\npaste this last:\n\n> from now on run this on a schedule.\n\n> \n\n> - daily: pull new installers in the regions i'm working and add them to the list\n\n> - daily: run the render queue for any batch i've queued and put the results up for me to check\n\n> - weekly: tell me which campaigns have gone quiet on enquiries and what's still in the print queue\n\n> - monthly: what each installer paid, what each campaign cost, and what's left in the print budget\n\n> - if nothing moved, say that in one line rather than filling the report\n\nwhat can break\n\nstreet view imagery is old in a lot of places. you'll get houses with the wrong colour paint, scaffolding etc., this is the single biggest cause of a card that looks wrong.\n\nthe metadata endpoint is how you catch most of it, since it gives you the date the photo was taken before you pay for the image.\n\nwhen a render looks off, the answer is usually not enough reference. one bad street view frame and the model invents the rest of the house. pull more angles for that address and run it again.\n\naddress data might not be right. validate before you print, because a card to a dead address will still cost you.\n\nwhatever you're selling the installers, the conversation has to happen in september.\n\nand start building this now. the season books in the next six weeks.\n\nhope you have fun setting this up.\n\njoin here for more value: https://t.me/+pbCBBtUEtu1lZDA1"},"adhxContext":{"savedByCount":1,"publicTags":[],"previewUrl":"https://adhx.com/everestchris6/status/2093378281810833789"}}