Curiosity Technology | Inspired by the future

What we build.

We build AI products that work in the physical and digital world and interact with people.

Explore Solutions

Physical & Digital AI Assistants

Visit

AI that talks, understands and responds in real time.

INPUTVoice & Touch RecognitionReal-timeCOMPUTEEdge AI InferenceOn-deviceOUTPUTMultilingual Response24 / 7

AI Hardware

We design and manufacture the devices our AI assistants run on.

Our devices are in 5 countries

Our reach

2.5M+

People reached

AI Models

AI models built for your needs, able to run on your own infrastructure too.

Open-source AI

We share the models and technologies we build with the community.

Explore on Hugging Face ↗

Technology

At every stage Advanced technology, Deep engineering

We ship strength at every layer of the stack with models we design and train ourselves.

LLM

Mars

Purpose-built language model tuned for Turkish structure and cultural context. Open source on Hugging Face.

EDGE AI

Moon

A minimal model that runs on a CPU. Designed for seamless human interaction on IoT devices.

MULTIMODAL LLM

Venus

A multimodal LLM built as an omni model: it understands text, images and audio together, in a single model.

AUDIO AI

Voice Isolation

Separates speech from background noise. Designed for low-latency conversational experiences.

AUDIO AI

Noise Cancellation

Clear isolation in loud environments, airports, malls, and busy venues.

AUDIO AI

Turn Detection

Detects who should speak next in real time. A core primitive for natural dialogue.

AUDIO AI

Uran

Speech recognition model that turns speech into text, optimised for edge systems.

PRIVACY

Runeward

A protection layer for AI interactions that detects and masks personal data.

LIP-SYNC

Phoenix

Efficient lip-sync engine that runs on a CPU, with no GPU needed. Matches speech to natural mouth movements.

View all models on Hugging Face

Planet surface maps: Solar System Scope (CC BY 4.0) and NASA

Enterprise AI infrastructure

Powerful AI. Full control.

On-premise

Your data stays on your infrastructure.

Guardrails

You set the rules.

Low latency

Real-time interaction.

High throughput

High processing capacity.

The power is yours. So is the control.

Global Impact

Global scale,
local rules.

Our systems run on site, with data residency controls, compliance-ready deployments and sub-150ms inference at every node.

Countries

Continents

<150ms

Avg latency

99.9%

Uptime SLA

Shape the Future with Our AI Solutions

Our Products

AVATAR AI

SelamAI

Interactive AI kiosk platform with real-time avatars and human-level conversations. Built for malls, hotels, hospitals and other high-traffic venues, multilingual, sub-second latency, and beautifully reliable.

Learn More

AIRPORT AI

PortAssist

Physical AI assistant built for airports. Real-time avatar kiosks guide passengers to gates, transfers and services in their own language, designed around airport operations and passenger flow.

Learn More

AI FOR EVERYONE

ZekAI

An AI-powered mind map. Every thought has a pattern beneath it; when mind maps meet AI, innovation unlocks. Free for everyone.

Learn More

ON-PREM AI

SezAI

On-premises ChatGPT alternative for enterprises. Run advanced generative AI fully inside your private network, no data leaves your perimeter, with full audit, governance, and access control.

Learn More