2026AI · Web · Mobile — now booking Q3

AI-powered software
shipped from Romania.

Romania-based software studio with deep experience in production AI — LLM agents, RAG, copilots and AI-native product workflows. We also ship full-stack web and mobile for founders and growing teams.

What we do

AI, web & mobile development services.

Production AI — copilots, RAG and agents — plus web, mobile and SaaS platforms from the same team, end-to-end.

01

Production-grade React, Next.js, Node and PHP platforms — from MVP to multi-tenant SaaS.

02

Swift, Kotlin and React Native apps that ship to the stores and stay there.

03

Our fastest-growing track: LLM agents, RAG, vector search, streaming copilots and AI image pipelines — shipped inside live products.

04

TypeScript, Go and Python services. Postgres, Redis, queues and observability — the whole spine.

05

Design that compounds. Tokens, components and motion principles — handed over ready to ship.

06

Embedded with your team. We co-own roadmaps, ship weekly, and stay until it works.

Live from the keyboard

This is how we ship.

Production code for AI agents, web platforms and mobile apps — with live previews of what we ship.

app/Hero.tsx
api/agent.ts
mobile/Login.swift
TSX
EXPLORER
▾ app/
◇ Hero.tsx
◇ Footer.tsx
▾ api/
◇ agent.ts
◇ auth.ts
▾ mobile/
◇ Login.swift
◇ Home.swift
1
Live preview
We build software that ships.
← magnetic
✓ compiled
main
app/Hero.tsx
Ln 1Col 1UTF-8TSX
Selected work

Recent work across software and AI.

A small slice of shipped product work. Public case studies are shared where client approval allows it.

E-commerce / Web

Todome

Multilingual, multi-country e-commerce platform for gates, fences and gate automation. Symfony ERP and API, Next.js storefront and React backoffice.

Private engagement
SaaS / Multi-tenant

Churcheos

Multi-tenant SaaS for church management with billing, locale support and queue-driven notifications.

Private engagement
AI / SaaS

Scena

AI product photography workflow for e-commerce teams, with batch processing and marketplace integrations.

Case study on request
AI / E-commerce

AutenticGift

AI-powered photo personalisation for custom gifts, including background removal and portrait editing workflows.

Private engagement
AI / Voice & chat agents

Agent Platform

Voice and chat agents grounded in an approved knowledge base, with real tool execution, provenance-aware memory and a built-in simulation and QA harness.

Internal product
AI / Desktop

Transcribe

Local-first transcription and AI summaries for work conversations. Whisper on the user’s own GPU, crash-proof recording, nothing leaves the machine.

Internal product
AI / Internal tools

Estimate Builder

AI-assisted pre-sales workbench that turns a client brief into a traceable, gate-checked proposal: estimate, PRD, SOW and delivery backlog.

Internal product
Tools of the trade

A stack chosen for production, not for novelty.

TypeScript
React
Next.js
Node.js
Python
Go
PHP
Laravel
Symfony
Swift
Kotlin
React Native
Postgres
Redis
Kafka
AWS
GCP
Vercel
OpenAI
Anthropic
LangChain
TS
RN
GO
AI
PG
AWS
NX
KT
How we work

A short, opinionated process.

Four phases. No theatre.

01

Discover

A week of deep listening. We map the problem, the users, and the constraints — no slides.

02

Shape

Wireframes, prototypes, a technical spike. Decisions, not opinions.

03

Build

Two-week sprints. Shippable increments. Live demos every Friday.

04

Launch

Hardening, observability, runbook. We stay on call through the first month.

15+
Shipped projects
50%+
Engagements with AI
4
Core delivery tracks
RO
Romania-based studio
Questions

Things people often ask.

Yes — AI is a core delivery track for us, not an add-on. We ship LLM agents, RAG systems, vector search, streaming copilots, and AI image workflows in production for founders and product teams.
OpenAI, Anthropic, LangChain, pgvector, Redis, and custom TypeScript/Python services. We pick models and infrastructure for latency, cost, and reliability — not hype.
Often, yes. We usually start with a focused spike: retrieval quality, prompt design, evals and a production-ready vertical slice wired into your current stack — not a sandbox demo.
Most projects kick off within 7–10 days. Smaller scopes can begin the same week if there's urgency.
Both. Fixed prices for discovery and MVP projects; time & materials for ongoing product work where requirements evolve.
Yes. We usually start with a focused technical audit: architecture, code health, risks, and a practical delivery plan.
We are based in Romania and work with a trusted network of product, design and engineering collaborators when a project needs a larger team.
We usually start with a focused discovery or MVP scope, then scale into longer product work when there is a clear roadmap and delivery rhythm.
We are based in Craiova, Romania. Email hello@m-tech.dev or call +40 785 464 716 — we reply within one business day.
Yes. We work with clients across the European Union, the United States and further afield, in English or Romanian, in the Eastern European time zone (UTC+2/+3).
For agencies

You won the pitch. We ship under your brand.

White-label AI creatives, campaign sites and e-com tooling. You keep the client — and the margin.

Have something worth building?

Tell us about it. We'll respond within one business day with an honest assessment of whether we're a good fit.