Bot-marketing partner edition · EN
🌍 For Partners
Partner shaking hands with Alina, bots being assembled the same day
← Full catalog (RU)
Partner Edition

Agents in hours.
Not days.

This is the partner playbook for Bot-marketing: run our constructor on your own server, build GPT agents for your clients with MCP + Claude Code, and grow with advertising and referral machines we have already battle-tested on ourselves. Four scenarios: how you build, how you sell, how you scale — and how we launch this together.

Meet Mario — an agency owner who became our first partner. His story is your onboarding. ↓

The Cast
Mario
Mario · Partner
Digital agency owner. Ships chatbots to his clients — now in hours, on his own infrastructure.
Feels: ambition — “my market, my server, my margin.”
Alina Kobrova
Alina Kobrova
The face of Bot-marketing. For partners she is the guide: setup, migration checklists, referral math.
Feels: excitement — “new country, same cobra on the shirt.”
Terminalka
Terminalka
The silent executor: Claude Code sessions that assemble agents through the MCP server.
Feels: nothing. Builds bots in any language with equal silence.
Scenario 1 · The Workshop
Monday, 2 PM. A client brief arrives. Monday, 6 PM — the agent is live.
Developer builds agents via MCP and Claude Code in hours
Hours!
MCP + Claude Code

Build GPT agents in an afternoon

How it works: your developer opens Claude Code and connects to the Bot-marketing constructor through our MCP server. Then they simply describe what the client needs — “a booking bot for a dental clinic, three branches, Google Sheets as the schedule” — and Claude Code assembles the flows, connects the channels and runs test dialogs through MCP, the same way a human would click through the console. Only faster, and without typos.

What used to take days — flow design, block wiring, copy-pasting prompts — collapses into hours. Your senior reviews the result; juniors are optional.

The unit economics flip: an agent built in 4 hours can be sold at a day’s price — the margin is yours.
Scenario 2 · The Ad
The billboard says what nobody else dares to promise.
Speed-promise advertising brings leads, demos convert them
Leads!
Advertising that only works if it’s true

“Any chatbot in hours, not days”

The campaign: Mario runs ads with one message — we deliver any chatbot in hours, not days. Competitors can’t copy the claim, because they can’t copy the workshop behind it.

The funnel: every lead that comes in gets a free demo bot the same day — built through MCP from the lead’s website and social profiles, exactly like our own sales conveyor does it (see issue #5). The prospect doesn’t read a proposal; they chat with their own bot.

Free demos used to be too expensive to give away. At 30 minutes of machine time, they become your cheapest closer — and conversion jumps.
Scenario 3 · The Influencers
Every paying client pays the person who brought them.
Influencer integrations feeding the referral program
Ref!
Instagram + referral program

Influencers on performance, not on faith

The move: Mario buys ad integrations from Instagram influencers — but instead of paying only for reach, he plugs every influencer into a referral program: a personal link, tracked leads, and a bonus for every client who actually pays.

Our part: we help set it up on the Bot-marketing stack — the same referral machinery our own employees use (dossier T2): personal links, CRM attribution, automatic bonus accounting, payout reports. The influencer sees her leads and her money in real time.

Influencers stop being a lottery: reach is the ad, payouts are for results. The best ones become a sales channel, not a media buy.
Scenario 4 · The Launch Plan
Moving day: the constructor gets a second home.
Migration of the constructor and services to the partner's server
Launch!
How we start — step by step

Your server, our stack

First, on our side: we rework the constructor for partner deployments — every module becomes multilingual and updatable, so your copy can speak your market’s language and receive upgrades.

Then, on your side: you rent a server and own its uptime. You pay us for the migration work and for help porting existing projects from the old server to the new one. Our current financial agreement stays exactly as it is — nothing changes in how you pay us.

The license is honest: your copy is bound to your hardware, updated yearly, with its own limits. Your infrastructure — your control; our updates — your evolution.
🧩 Constructoryour own copy: multilingual modules, yearly updates, hardware-bound license
↓ ships together with ↓
🔐 OpenAI Proxyyour token vault: LLM keys and third-party service tokens (e.g. Google Sheets) that modules use
🔌 MCP Serverthe door for Claude Code: build and configure agents in the constructor programmatically
🖥 Consolethe professional cockpit: configure client scenarios, manage users, run your operation
We · Bot-marketing
Rework the constructor: multilingual, updatable modules ready for partner deployment.
Migrate the stack to your server: constructor + OpenAI Proxy + MCP Server + Console.
Help port existing projects from the old server to the new one.
Ship yearly updates for your copy.
Help configure the referral program and share our ad playbooks.
You · Partner
Rent the server and keep it running — it’s your infrastructure.
Pay for the migration and porting work; financial terms otherwise stay unchanged.
Keep your LLM and service tokens in the OpenAI Proxy vault.
Build client agents through MCP + Claude Code; configure scenarios in the Console.
Run the growth machines: speed-promise ads, free demos, influencer referrals.
Token Economics
$10,000
earned on tokens alone last month — by us, on our own installation

Every agent burns LLM tokens, and all of them flow through the OpenAI Proxy on your server. That makes tokens a revenue line, not a cost line: we add a shared margin on top of the provider price — invisible to the end client, standard for the industry.

Clients pay for tokens as part of the service — the way they already expect to.
The margin on tokens is split between us the same way we split subscription revenue — same proportions, no new math.
The more agents you ship, the more this line grows: it scales with usage, not with your hours.
The Price Tag

Your own constructor, the migration, OpenAI Proxy + MCP Server + Console, porting help, yearly updates, the referral machinery and every playbook on this page. What does the whole thing cost?

$ ???