01
Remembers approved corrections
You approve a rule or correct a result, and we add it to the agent's instructions. The next run follows the agreed process.
// AI AGENTS · LAUNCH IN 7 DAYSYour AI co-worker.
Your AI co-worker.
Takes the routine. Reports in Telegram.
We launch one clear workflow in 7 days — from reports and documents to support. Launch is 49,990 ₽, with the first month of support included.
// 01 · CHAT vs AGENT
The difference is not the model. It is who does the work.
Often it is the same model inside. The difference: a chat answers when asked. An agent acts on its own, on schedule, inside your tools.
| Parameter | ChatGPT | Your agent |
|---|---|---|
| Remembers you tomorrow | Starts from zero | Remembers across chats |
| What it does | Answers with text | Performs actions |
| On a schedule | Waits for a prompt | Summaries, reports, reminders on its own |
| Where it lives | A separate tab | In your Telegram |
| Your data | On someone else's servers | CRM, sheets, email; you revoke access |
// 02 · CAPABILITIES
It does more than answer. It follows your rules.
We start with one measurable workflow. After the pilot, we connect more tools and tasks.
02
Five kinds of memory
The conversation, your rules, what you discussed a month ago, what's in your documents — and how to do your tasks.
03
Runs on schedule by itself
The 7:30 summary, the Monday report, a price check every four hours. Without you.
04
Actions, not advice
Creates tasks, answers clients, prepares files. Through MCP tools.
05
Voice
Dictate in Telegram, get it done.
06
Search with sources
Fresh data and a summary. Sources you can check.
// 03 · STACK & RELIABILITY
Engineering, not a wrapper over a bot.
GPT-5, Claude and open models. Chosen for the task and budget.
Servers in Russia, managed by us. Revoke the agent's access anytime; the individual setup can run on your own server.
Safe start: for the first weeks the agent drafts, you confirm. Autonomy grows gradually.
Payment in rubles. No foreign cards or keys, we handle the servers.
OUR OWN STACK
Наш стек: Claude Code, Codex, opencode, Hermes, OpenClaw, Paperclip.
The stack we work in every day: Claude Code, Codex, opencode, Hermes, OpenClaw, Paperclip.
// 04 · PROCESS
Three steps. We set it up.
01
PILOT · METRIC FIRST
We take one narrow slice, fix a metric before we start (e.g. client reply time). First result within a week.
02
BUILD · WE DO IT
We assemble the agent: memory, actions, integrations, server. First result within a week.
03
SUPPORT · ONGOING
We keep it running: fixes when external services change, scenario updates. An agent with no one to fix it lives until the first failure.
// 05 · PRICING
Two ways to start.
We do not automate everything at once. We take one repeatable task, define the metric, and show a working result in the first week.
Starter agent
0,990 ₽launch
0,800 ₽ / monthfirst month included
- Lives in your Telegram, on our server in Russia. Payment in roubles.
- It remembers you. Your rules, your numbers, what you discussed a month ago — no conversation starts from scratch.
- Follows approved rules. Your corrections are added to the agent's instructions.
- Scheduling in plain words. «Every morning at 7:30 — the summary.» No developer needed.
- Voice. Dictate in Telegram, get the result.
- Web search with links to the sources.
- NOT included: your email, drive, documents, CRM — that's a separate access layer, see the package below.
Built around you
from0,000 ₽scoped to the job
- Everything in the base package.
- Full memory: it knows you personally, remembers your rules and past decisions, and searches your documents by meaning — not by keywords.
- Deep research: drives a browser, gathers and cross-checks sources, writes the analysis — research, not one search result.
- Your services: email, drive, Google Docs, calendar. It reads them and drafts replies itself.
- Your systems: CRM, marketplaces, accounting, databases and internal APIs.
- Autonomous work: monitors, summarizes, handles inbound — without you, by your rules.
- Isolated execution with an audit log: you see what it did and why.
// 06 · TASKS
What it takes off your plate.
01/09
Support
Clients write at night, you can't keep up.
Replies from your templates; you approve with one tap.
>watch(inbox) -> reply(template)
02/09
Email & docs
The inbox piles up; the contract was needed yesterday.
Sorts email, drafts replies, edits Google Docs and Sheets.
>sort(mail) -> draft(reply)
03/09
Decks
You build proposals by hand in the evenings.
Assembles the deck and the proposal from your template — in minutes.
>build(deck) -> export(pdf)
04/09
Video
Reels and Shorts due every week — you never get to them.
Writes scripts, cuts clips, builds covers and titles per platform.
>cut(reels) -> caption() -> publish()
05/09
SMM · all channels
Telegram, VK, YouTube — no one to run them all at once.
Runs Telegram, VK, YouTube: content plan, posts, replies — one brand voice.
>post(tg, vk, yt) -> schedule()
06/09
Code
Small fixes wait weeks for a developer.
Writes scripts, repairs integrations, automates routine in code.
>fix(bug) -> deploy()
07/09
Reporting
Sunday evening goes to the summary again.
The team report is ready by Monday morning. By itself.
>cron("Sun 20:00") -> report()
08/09
Sales
Forgot to call back — the deal went cold.
Remembers clients, reminds you to call, briefs you before each one.
>remind(call) -> brief(client)
09/09
Marketplaces
A competitor cut prices; you found out three days later.
Morning digest: sales, stock, competitor prices on your platforms.
>watch(wb, ozon) -> alert(price_drop)
// 07 · QUESTIONS
Answering what gives you pause.
You can, just like installing a CRM yourself. The question is who fixes it when Telegram updates or a service changes its API. We build and maintain.
On servers in Russia. The starter plan runs on infrastructure we manage; the individual setup can run on your own server. Access is revocable anytime.
The starter agent is 49,990 ₽ to deploy one working workflow. The first month of support is included; after that it is 9,800 ₽ a month for the server, models, support, and API maintenance. Custom process work starts at 150,000 ₽; we quote it after a twenty-minute review.
First scenarios within a week. Then we expand.
It won't replace a human where judgement and accountability are needed. It doesn't make phone calls. It can't touch anything you haven't given it access to. When unsure, it escalates to you instead of inventing an answer.
For the first weeks it only drafts. Correct it with a sentence, it remembers the fix.
At low confidence the agent escalates to a human instead of making things up. Behavior is testable before launch: you approve actions, you do not take them on faith.
No. The code and the agent are yours, models are interchangeable, infrastructure stays under your control.
// 08 · CONTACT
Start with one task.
In 20 minutes, we will define the first workflow, success metric, and pilot boundaries. No long brief or obligation.
First workflow in one week.