Outbound end to end
The nine documented stages: context, infrastructure, lists, copy, approval, launch, replies, measurement and deliverability. The same process we run for our clients, every stage written down with its deliverable.
Outbound, content, copy and pipeline. The same software we operate, on your credentials and your infrastructure. It’s not a course or a framework. It’s the software.
GTM OS · how it works
Seven concrete things, each with what you do and what you save. The platforms it’s built with come next.
The nine documented stages: context, infrastructure, lists, copy, approval, launch, replies, measurement and deliverability. The same process we run for our clients, every stage written down with its deliverable.
Pillars per profile, learned voice, idea bank, design with your brand kit and OAuth publishing. The pieces we publish ourselves come out of here: the text and the image.
SEO and AEO: llms.txt, JSON-LD and structured content for search engines and for models. More and more people ask a chat instead of searching, and there the question isn’t whether you rank, it’s whether you get cited.
Subscribers, editions, sending and metrics, running on your domain. Without paying a third party per subscriber.
Templates ready to publish, with structure, design and technical SEO solved. What changes is your content, not the skeleton.
On Meta’s official API, with your own Business Portfolio, and only with people who replied and gave you their number and their OK. It connects by API to the flow your company already runs: the tool adapts to your processes, not the other way around.
Copywriter, reply manager, domain health, campaign analyzer and copy review. They aren’t on GitHub: they exist because a client needed them and nothing else did the job.
The difference between this and a GitHub fork is that here the variable cost is yours because the keys are yours, and the maintenance is ours because the software is the same one we operate every day for our clients.
One-to-one copy
It resolves every block of the message lead by lead against your library and your dimensions, with auto-QA before export.
You don’t write variants by hand or send the same email to two different roles.
Runs on your Anthropic key and your Supabase
LinkedIn content
Pillars per profile, learned voice, design with your brand kit and OAuth publishing.
You don’t coordinate a writer, a designer and whoever publishes.
Runs on your Anthropic, your Perplexity and your LinkedIn
Per-number pacing, cross-channel suppression, inbox, and Meta’s limits respected by design.
You don’t burn phone numbers or fight the API.
Runs on your own Meta Business Portfolio and WABA
Reply Manager
Unified inbox for email, LinkedIn and WhatsApp, with AI drafts and human sending.
You don’t check dozens of mailboxes.
Runs on your Smartlead, HeyReach and Meta accounts
Domain health
Daily score per domain against its campaign cohort, and silent-mailbox detection.
You don’t find out a domain burned once nobody replies.
Runs on your domains
Analyzer
Which role, country, company size and wording converted, in your own cohort.
You don’t decide on a benchmark from a blog post.
Runs on your Smartlead and your HeyReach
Reporting and pipeline
Unified pipeline and dashboard from day one, across the channels you use.
You don’t build the report by hand or keep pipeline in three places.
Runs on your API keys
Copy review
Versioning, comments anchored to the text and recorded approval.
You don’t approve copy over an email chain.
Runs on your Supabase
Nobody fills in sixty fields. The OS researches your company and hands you the structure back, and you correct what’s wrong.
Two inputs. The domain, and whether you sell into LATAM or the United States. The second one sizes everything that follows.
The OS researches your company, your category and your competitors, and synthesizes what it finds.
That produces your company types and buyer personas, each with its attributes: value prop, pains, cases, positioning. You correct whatever doesn’t fit.
You connect your accounts: Anthropic, Smartlead, HeyReach, Meta, LinkedIn, your Supabase. The variable cost is yours because the keys are yours.
Every module reads your company/ before doing anything, so copy, content and routing come out with your context, not a generic one.
Pipeline and dashboard exist before the first message goes out, not on month three.
Three years of running outbound for 50+ companies, available from day one. You don’t start from zero. You start from here.
What you get access to
You start with a list: 1.4M leads and 970K companies
And with the record of who’s already been contacted, so you don’t send the same email to someone who got it last month from another campaign.
You know what converts before sending the first email
Which role, which country, which company size, which sequence step and which wording. Per vertical. It’s what the Analyzer learned from three years of sending across fourteen clients, not a benchmark from a blog post.
You don’t pay twice for the same data
If we already enriched part of your list, you don’t pay Apollo or Findymail for it again. On a typical list that’s a large part.
You know whether your deliverability is good or bad
Your numbers are compared against roughly 1,000 domains and 2,500 mailboxes we monitor. A 96% on its own says nothing; against that base, it does.
You know how much infrastructure you need for the meetings you want
Depending on whether you sell into LATAM or the United States. It comes from operating both markets, and it’s the first question of onboarding.
All of this is queried from inside the OS. It isn’t exported or handed over: it’s the asset, and it’s why the GTM OS costs what it costs and not zero.
Runs inside your company
The eight platforms, on your credentials and your infrastructure. The software is yours while you use it, and so is the variable cost, because the keys are yours.
We don’t do
B2C lead generation
The system is built to sell to companies: ICP, job titles, buying signals, domains. None of that exists when the buyer is a person.
Paid media (work in progress)
We’re building it. We won’t sell it until we can measure it with the same precision as the rest of the system.
The difference isn’t how much software you get. It’s the same. The difference is who gets it running the first time.
$400/mo
Guaranteed: if you miss the goal, we support you for freeIf you have someone technical who can give it a few hours a week.
On request
Guaranteed: if you miss the goal, we support you for freeIf you want the system in-house but don’t want to set it up yourself.
Everything starts with a call. There’s no buy button on purpose: onboarding is calibrated to the market you sell into and the modules you need, and that doesn’t get solved in a checkout.
It’s the answer to the question we get most from people who already hired agencies that overpromised: how do I buy this without a leap of faith. It isn’t a line on a website, it’s a clause.
01
We do the math together
On the first call we start from your business (deal size, close rate, sales cycle, how many meetings you need for the numbers to work) and from the plan you pick, which sets the volume that goes out each month. Those two set the goal for each month.
02
We sign it
The monthly goal and the term, usually three or six months, are written into the contract. Not in a proposal or an email.
03
If the goal is missed
Your team runs the system, so what we put in is free consulting and support on your operation, until your team reaches the goal.
We count qualified meetings: ones that happened and that fit the ICP we defined together. We don’t promise closed deals, because those depend on your sales process. And there’s no number here because the goal is built from your numbers and the plan you pick, not another client’s.
Chatting with Claude Code connected to the GTM OS, which runs our tools, processes and learnings behind it. You ask for the list of companies in a country, the sequence, or to load the batch, and the system runs each stage with your company’s context behind it. What’s worth checking by hand is little and always the same: the copy before it goes out, the replies in the unified inbox and the health of the domains. That’s what the portal is for: pipeline, metrics per campaign and the status of each channel in one place, without going tool by tool.
The first 3 to 4 weeks are the heavy ones: buying domains, creating and warming up the mailboxes, running the deep research and building the first list and the first sequence. That’s when it takes the most time. After that it drops to a few hours a week: answering replies from the inbox, approving the copy for each batch, loading new leads and checking the metrics once a week. If that’s more time than you have, the mode we operate suits you better and we’ll say so.
They’re on top of the $400 for the GTM OS and you pay them on your own accounts. The math starts with how many meetings you want per month. In our email campaigns to LatAm, one meeting is booked every 347 leads contacted, about 1,500 messages. That sets the infrastructure: each mailbox sends 20 emails a day and each domain carries 2 mailboxes, so one domain enables 40 emails a day. A domain costs about USD 14 and is rotated every 3 to 6 months; a mailbox, USD 3 to 3.5 a month. Adding sending, mailboxes, AI for the copy, contact data and verification, 5 meetings a month run about USD 250 to 400 in tools; 10, about USD 350 to 600; 20, about USD 650 to 1,000. Domains are separate because they aren’t monthly: for 10 meetings, about USD 252 every 3 to 6 months. LinkedIn is optional and adds per account, and the United States takes about five times the infrastructure. We tune it with your numbers at kick-off. You paying for them is deliberate: if the variable cost were ours, $400 wouldn’t be a real price, and what runs stays yours even if we stop working together.
| Meetings per month | Leads per month | Emails per day | Domains / mailboxes | Domains, every 3 to 6 months | Tools per month |
|---|---|---|---|---|---|
| 5 | 1,750 | 340 | 9 / 18 | USD 126 | USD 250 to 400 |
| 10 | 3,500 | 680 | 18 / 36 | USD 252 | USD 350 to 600 |
| 20 | 7,000 | 1,360 | 35 / 70 | USD 490 | USD 650 to 1,000 |
LatAm, email only, list prices as of October 2026. LinkedIn and Clay are extra.
Partly. The skills that don’t call a vendor run on Claude Code alone: qualifying companies against your ICP, mapping people by title, pulling signals from websites, checking your domains’ DNS and the deep research on your company. But what makes the system perform is fed by the tools: the copywriter writes with the Anthropic API, and the portal, metrics per campaign, unified inbox, Analyzer and domain monitor read from the sending platform. Without them you have the process, but not the operation or the visibility. That’s why we recommend paying for them from month one. For 10 meetings a month in LatAm, the stack comes to about USD 350 to 600 a month, plus domains.
| Tool | What for | Without it | USD per month (10 meetings) |
|---|---|---|---|
| Claude Code | Day to day: you run the system from it | No GTM OS | from 20 |
| Smartlead | Email sending | No email dashboard, unified inbox, Analyzer or domain monitor | 94 |
| Domains and mailboxes | Where the email is sent from | No sending | 108 to 126, plus 252 in domains every 3 to 6 months |
| Anthropic API | Copywriter, reply drafts and Analyzer | No one-to-one copy or campaign analysis | 30 to 70 |
| Apollo, Findymail o BetterContact | The decision maker’s email | You bring the list with emails | 100 to 250 |
| BounceBan | Verifying before sending | More bounces and burned domains | 20 to 40 |
| HeyReach (optional) | No LinkedIn or its metrics | 63 to 79 per account | |
| Clay (optional) | Orchestrating enrichment and signals | The skills cover it | from 185 |
List prices as of October 2026. The tools are examples: if you already pay for others that do the same, we check at kick-off whether they connect.
Yes, with an honest caveat. Our own database is built on what our clients sold: SaaS, fintech and B2B services. If your ICP is different, say clinics or diagnostic imaging centers, it isn’t fully there. The way in is scraping Google Maps and company websites country by country, with third-party tools or the scrapers we already built, and then finding the decision maker (owner, director or administrator) in our database and other sources. We set it up together at kick-off.
No. It’s monthly and you cancel whenever you want. Since everything runs on your accounts, what you built stays in your company.
Yes, and it’s in the contract. On the first call we start from your business (deal size, close rate, sales cycle, how many meetings you need for the numbers to work) and from the plan you pick, which sets the volume that goes out each month. Those two set the goal for each month. The monthly goal and the term, usually three or six months, are written into the contract. Not in a proposal or an email. Your team runs the system, so what we put in is free consulting and support on your operation, until your team reaches the goal. We count qualified meetings: ones that happened and that fit the ICP we defined together. We don’t promise closed deals, because those depend on your sales process. And there’s no number here because the goal is built from your numbers and the plan you pick, not another client’s.
You’re not left alone with the software. You get the community, with moderators from our team who support the process, and the updates: the system is the same one we operate every day for clients, so what we learn in their campaigns reaches you.
You can, and for testing it’s fine. What you won’t have is the infrastructure that sustains it or three years of campaign data. A skill published by a stranger gives you the process once; this is the software we operate every day for paying clients, and it stays maintained because we maintain it for them.
No. That database, the priors on what converts per vertical, the enrichment cache and the deliverability benchmarks aren’t handed over: they’re queried. They’re three years of sending across fourteen clients and they’re the asset. We say it here instead of in the fine print because it’s the question you were going to ask on the call anyway.
It replaces the part of the work that’s building the system, which is where the first three to six months of a hire go. It doesn’t replace judgment: someone on your team still has to decide who you write to and approve what goes out.
You can switch at any point, and the other way around too. The modules are the same on both sides; the only thing that changes is who operates them. Plenty of people start with us operating and take it in-house once they have someone to put on it.
Because onboarding is calibrated to the market you sell into and the modules you need, and that doesn’t get solved in a checkout. We would rather have a twenty-minute call than a client who bought the wrong configuration.
This is why we work with enterprise and VC-backed companies.
Book a Discovery Call