Technologies
Technologies we work with
The chips, frameworks and AI tools behind our client projects.
We choose technology to fit the product, not the other way round. These are the platforms and tools our team uses most in client work.
Embedded and hardware
- Microcontrollers: ESP32, ESP32-S3, nRF52832, nRF52840, STM32.
- Embedded Linux and edge AI: Raspberry Pi, BeagleBone, Rockchip RK3576 and RK3588.
- Firmware: C and C++, ESP-IDF, nRF Connect SDK, FreeRTOS, Zephyr, bare-metal.
- Connectivity: Bluetooth LE, Wi-Fi, LoRa, GSM, RS-485, Ethernet, USB.
- PCB: KiCad, 2–4 layer and flexible PCBs, custom carrier boards.
AI and automation
- Models: OpenAI GPT, Anthropic Claude, Google Gemini and local models where data must stay on site.
- Automation: n8n, webhooks, Google Workspace and CRM integrations.
- Channels: WhatsApp Business Platform, website chat, email.
- Patterns: AI agents with tools, retrieval-augmented generation (RAG), structured outputs and human review.
Apps and backends
- Mobile: Flutter and React Native for iOS and Android.
- Web: Angular, React and modern JavaScript.
- Backend and cloud: Node.js, Laravel, Python, REST APIs and cloud hosting.
How we choose
We start from power, cost, connectivity, data and maintenance needs. Our guides on nRF52 vs ESP32, Zephyr vs FreeRTOS and n8n vs Zapier show the trade-offs we discuss with clients.
Frequently asked questions
Do you work with chips not listed here?
Often, yes. Firmware skills transfer across Arm Cortex-M and similar MCUs; we confirm in the scoping call.
Which AI model do you use?
The one that fits cost, accuracy and data rules for your case. We regularly use GPT, Claude and Gemini, and local models when data must stay on site.