Alljoined is a neurotechnology startup building foundation models for brain data, focused on decoding visual perception from non-invasive EEG signals. The company develops AI systems that reconstruct seen images from EEG recordings, working toward scalable, generalizable brain-computer interface applications.

The company's technical contribution centers on large-scale EEG-to-image datasets and multi-subject decoding models. Its ENIGMA model learns a shared "visual language" across different brains, enabling rapid adaptation to new users in minutes rather than requiring hours of individual calibration data. The underlying datasets capture evoked brain responses to thousands of natural images using 64-channel EEG, with careful experimental design to ensure signal quality including response-based stimulus timing, cross-session repetition, and diverse image classes.

Alljoined targets the intersection of neuroscience research and consumer BCI development, providing tools to better understand visual processing mechanisms and advance image reconstruction from non-invasive neural recordings. The company's datasets and models are publicly released to support the broader research community. Applications span clinical BCI contexts (communication aids for locked-in patients) and eventual consumer use cases for thought-to-image translation.