RU

SERVICES & PRICING

What I do and what it costs

From web development and long-term partnership to AI integration and one-off tasks. Indicative pricing, contract, receipts.

Long-term partnership core offering

Technical partner on retainer

I cover the full IT side of your business as an in-house engineer would: development, infrastructure, automation, support — without the headcount cost.

from 50,000 ₽/mo

fixed-hour pool or incident-driven SLA — shaped to the workload

  • Web development and improvements to existing systems
  • Servers, deploy, monitoring, backups, SSL
  • Integrations with 1С / Bitrix / CRM / marketplaces
  • Routine process automation (parsing, reporting, content)
  • Same-day reply on weekdays, communication via Telegram/email
Discuss partnership
Web development fixed price

Site, landing, fullstack app

Full cycle — structure, copy, design, code, deploy, analytics. From a one-day landing to a catalog site with CMS.

from 15,000 ₽

landing from 15k, corporate site from 40k, fullstack catalog from 90k

  • Astro / Next.js / React / TypeScript for modern sites
  • WordPress / Bitrix — for enterprise storefronts and integrations
  • Responsive + Lighthouse 95+ + complete schema.org
  • Deploy to Timeweb / Cloudflare / client VPS
  • Full handover of access — no vendor lock-in
Learn more
AI integration specialization

I turn a base LLM into a working tool for your task

RAG systems, multi-agent orchestration, LLM integration into business processes. Not "wire ChatGPT to a button" — but designing the context so the model actually closes a specific operation: support, content pipeline, request handling, agents inside an existing system.

from 40,000 ₽

PoC RAG / agent (1–2 weeks) — from 40k, full integration with memory and tools — from 120k. Long-term agent-system maintenance discussed separately.

  • RAG: indexing internal knowledge (docs, FAQ, tickets), retrieval, generation with source citations
  • Multi-agent systems: orchestrator + specialised agents, call graph, shared memory
  • LLM memory systems: persistence layer, context window as a resource, layered prompts
  • Anthropic SDK, Claude Code, OpenAI GPT API, MCP servers, tool use
  • Integration into existing systems: 1С, Bitrix, web apps, Telegram, internal APIs
Learn more
Automation & integrations frequently requested

Bots, parsers, API integrations

Telegram bots for notifications and orders, competitor catalog parsing, 1С ↔ site ↔ marketplace sync, Excel processing.

from 15,000 ₽

simple notifier bot from 15k, fullstack bot with admin panel from 50k, parser from 12k

  • Telegram Bot API (Python / Node.js)
  • Parsing (Playwright, BeautifulSoup, requests)
  • REST API: 1С-Bitrix, Bitrix24, YooKassa, Wildberries Seller API
  • Excel/CSV: pandas, openpyxl — bulk export and transformation
  • LLM-driven generation: descriptions, translation, rewriting
Learn more
IT support retainer

Personal IT specialist on retainer

I take the technical routine off your plate: sites, servers, domains, SSL, content, small fixes. "Your IT person on call" format.

from 15,000 ₽/mo

base package — up to 8 hours/month, overflow billed hourly

  • Content and price updates
  • Bug fixes, recovery from outages
  • Hosting, domain, SSL, backups, uptime monitoring
  • Reporting prep, ad-hoc analytics
  • Solution-choice consulting
Learn more
One-off task no retainer

Targeted hourly work

Improve someone else's project, fix a bug, set up analytics, do a one-time data export, give a consult.

from 1,000 ₽/hr

4-hour minimum; rate negotiable for ongoing partners

  • Improvements to existing projects on any stack
  • CRM and analytics setup (Metrica / GA4)
  • Marketplace product uploads
  • Consulting, code review, audits
  • Emergency hotfix — same-day response
Learn more

All prices are indicative — a starting "from". The final amount depends on project scope and is fixed in the contract once the task is discussed. A fiscal receipt for every payment.

How AI integration works in practice

The model itself is raw material. What turns it into a useful tool is the context: which data, which memory, which tools, which rules. I take a base LLM (Claude, GPT) and build everything around it that is needed for the model to solve a specific client task — not as a toy, but as part of a working process. 

  basic LLM ─┐
              ├──► context ──► working tool
 your task ─┘                 for a specific task

RAG system over internal knowledge

I index your documentation / FAQ / tickets / regulations, build relevance-ranked retrieval, plug in generation with source citations. The result is an internal assistant that answers questions for staff or clients and does not hallucinate, because it only answers from your base.

Case: tier-2 support of a large IT integrator. RAG system shipped in a week, auto-closed tier-2 requests for 2.5 months.

Agent for business process automation

Content parsing, rewriting and description generation, inbound request handling, email routing, product card copywriting from a template. The LLM does the work that used to take hours of human time, in the right tone and the right format.

Case: content pipeline of 22,000+ SKUs (1C + Bitrix24 + GPT copywriting + CDEK integrations) — primary project for the past two years.

LLM integration into an existing system

I embed LLM functionality into 1C, Bitrix, your web app or Telegram bot. Through REST or MCP servers. With quality control: where the model acts autonomously, where it asks an operator, where it logs the answer for review.

Stack: Anthropic SDK, OpenAI GPT API, Claude Code, MCP, my own context-management framework for agents (1.4 years of practice).

BUDGET CALCULATOR

Estimate your project cost

Pick your parameters — get an indicative budget range.

Project type
Number of pages / screens
Additional options
Indicative range: — Discuss the details →

Indicative. The final amount is fixed in the contract after the task is discussed.

FAQ

What stack do you work with?

Frontend: React, Next.js, Astro, Vue.js, TypeScript. Backend: Node.js, Python, PHP. Databases: PostgreSQL, MySQL, SQLite. DevOps: Linux, Nginx, Docker, PM2. For landing pages — pure static (max performance). For complex systems — fullstack apps with admin panels and APIs.

How long does the work take?

Landing — from 5 working days. Telegram bot — from 3 days. Catalog site or shop — from 2 weeks. Parser or automation — from 2 days. Exact timeline locked after the task is studied.

Will you take on existing projects?

Yes. I often pick up improvement, refactoring, fixes, and support of existing projects — even if someone else built them. I read other people's code, set up deploys, wire analytics, fix bugs.

Can you join my team as an external developer?

Yes. I can cover frontend, backend, DevOps, or automation as part of your team. Comfortable with Git, code review, task trackers (Jira, Trello, Linear). Communication via Telegram, Slack, or Zoom.

How does payment work?

50% upfront before work starts, the rest on acceptance. I accept card transfers and corporate bank transfers. A fiscal receipt issued via the "My Tax" app for every payment. For retainer support — monthly upfront billing.

How pricing works: the final number is fixed after the task is properly understood. I don't pad estimates — you pay for the work, not the agency markup.

Geography & paperwork: remote across Russia and CIS. Self-employed (NPD tax regime), contract with both legal entities and individuals, fiscal receipts via "My Tax". TIN 366606549484, status verifiable at  npd.nalog.ru .