Reka is an AI research and solution company. Our mission is to deliver best-in-class multimodal API for your use cases. We unlock the ability to understand the physical world, reason, and take action. Founded by a team of experts from Google DeepMind and Meta, we are backed with $170M from leading investors and strategic partners including NVIDIA and Snowflake.

We’re building systems that understand, imagine, and act within the physical world – models that learn from video, audio, text, and real-world signals, bringing perception, reasoning, and imagination into a single framework.

And the impact is already real.

In media, our systems tag and clip your assets with rich metadata, so moments become searchable and reusable forever. In security, they help officers work through hours of camera footage by letting them chat directly with their video library. In crisis response, they read a scene instantly, highlighting patterns and guiding responders toward where help is needed most, and in wearable devices, they power glasses that can see what you see, translating street signs or assisting as you move through the world.

Different models.
Different use cases.
One unified approach: Multimodal AI that don’t just process the world but to also to understand how the physical world works.