I build the boring, load-bearing part: error branches, retries, fallbacks and logging — so a failed run alerts you instead of silently doing nothing. Free scoping call, fixed price before we start.
Fixed scope, fixed price, documented handover.
Lead routing, CRM sync, document processing, reporting. Self-hosted or cloud — deployed, tested, documented.
Retrieval over your data, multi-step reasoning, tool use. With guardrails and an evaluation pass, not a demo prompt.
REST, webhooks, GraphQL, legacy systems with no SDK. Including the parts vendors call "not supported".
Telegram and web bots that ask the right questions, hand a structured lead to a human, and log every step.
Scrapers and pipelines that keep working: dedup, retries, proxy rotation, and a schema you can query.
A queue instead of a quote: unlimited workflow requests in priority order, monthly audit, direct line to me.
Problem → what I built → what it does now. Click to open.
Client-specific names, data and internal URLs are omitted on purpose. Numbers above are from the running systems, not projections — happy to walk through any of them on a call.
No open-ended hourly drift.
You describe the process, I tell you what is automatable and what is not worth it.
Scope, price and delivery date in writing before any work starts.
You see a running flow early, on your data — not a progress report.
Documented flows in your instance, 30 days of fixes included.
Price, speed, ownership, failure modes.
A single workflow with 1–3 integrations starts at $400. A multi-workflow system with an AI layer, logging and monitoring runs $1,500–4,500. Ongoing work is a retainer from $1,200/month. The scoping call is free and the quote is fixed before we start.
A single workflow: 3–7 days. A multi-workflow system or an AI agent with retrieval: 1–3 weeks. You get a working prototype early rather than a status update.
Yes. Flows live in your n8n instance, credentials stay in your accounts, documentation comes with the handover. No lock-in and no license from me.
That is the part most demos skip. Every flow ships with explicit error branches, retries with backoff, a fallback model where it matters, and logging that says what broke — so a failure alerts you instead of quietly returning nothing. I have measured that silent-failure rate on a live system and fixed it; ask me about it on the call.
Yes, and it is a normal setup: you own the client and the scoping, I build, test and document, your brand stays on everything. Usually starts with a small paid test project.
I work remotely, overlap with EU and US-East hours, English at C2. Async by default, calls when a call is genuinely faster.