# Meta's Brain2Qwerty v2 Decodes Typed Sentences From Brain Signals at 61% Word Accuracy

The non-invasive pipeline reads magnetoencephalography recordings, a large gain over prior surgery-free methods and a step toward assistive interfaces without implants.

- Published: 2026-07-19T05:30:22.062Z
- Canonical: https://polylog.news/ai/2026-07-19/meta-s-brain2qwerty-v2-decodes-typed-sentences-from-brain-si
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Meta AI](https://ai.meta.com/blog/brain2qwerty-brain-ai-human-communication/), [arXiv](https://arxiv.org/abs/2502.17480)

Meta detailed Brain2Qwerty, a deep learning system that decodes sentences people type from memory using only non-invasive brain recordings. The architecture combines a convolutional front end over short signal windows, a sentence-level tran…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-07-19/meta-s-brain2qwerty-v2-decodes-typed-sentences-from-brain-si (subscription information: https://polylog.news/pricing).