Meta just dropped the full code for Brain2Qwerty v2, turning raw MEG and EEG signals from people typing memorized sentences into real-time decoded text at the word and semantic level. The pipeline stacks a convolutional encoder with a transformer and character-level language model, lifting average word accuracy to 61 percent on data from nine volunteers and letting the best cases hit 78 percent with at most one error per sentence.
WHAT IS ON YOUR MIND?
We are testing the new Meta Brain2QWERTY open source platform with the Zero-Human Labs Neuron Decoder AI platform.
We are getting nearly full short sentences decoding in the garage with NeuroSky sensors taken from old toys!
More soon!
https://ai.meta.com/blog/brain2qwerty-brain-ai-human-communication/?utm_source=twitter&utm_medium=organic_social&utm_content=video&utm_campaign=brain2qwertyv2
Everything needed to train and run both v1 and v2 now sits on GitHub under a non-commercial license, so labs and independent researchers can replicate or extend the work without starting from scratch.
Performance scales with training volume, yet the v2 dataset remains under embargo and exact latency on everyday hardware is not reported, leaving open how quickly this moves beyond controlled typing tasks.
Positive users praise Meta's Brain2Qwerty v2 for real-time brain-to-text decoding as empowering for brain disorders and a strong open-source step, while negative users object over mental privacy risks and doubts about Meta.
No Digg Deeper questions have been answered for this story yet.
some exciting new work from our AI teams at Meta on non-invasive brain computer interfaces!
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.
Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.
We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.
🧵👇
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.
Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.
We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.
🧵👇
Highly recommended reading.
What an impressive use of LLMs and deep learning.
Achieves "real-time sentence decoding from non-invasive brain recordings, approaching levels of accuracy previously exclusive to techniques that require brain surgery."
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.
Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.
We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.
🧵👇
Meta doing cool shit and open sourcing it.
Hell yeah.
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.
Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.
We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.
🧵👇
Meta says Brain2Qwerty v2 can decode natural sentences from non-invasive brain recordings in real time, reaching 61% word accuracy.
The system was trained on about 22,000 sentences from 9 volunteers, each recorded for 10 hours with MEG while typing.
Meta compares that with 8% word accuracy from prior non-invasive methods. Its best participant reached 78%, with more than half of sentences decoded with one word error or less.
This is still controlled lab research: small participant pool, MEG hardware, active typing data, and company-reported results. Not a clinical communication device yet.
Meta is releasing the training code, while BCBL is releasing the v1 dataset, pushing brain-to-text research further into open neuroscience infrastructure.
I am so hyped for the future.
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.
Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.
We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.
🧵👇
Meta researchers used LLMs to predict the text a person was typing just from non-invasive brain recording!
Literal mind reading!
"For our best participant, the model accurately decodes half of the sentences with one word error or less." → INSANE
Let's learn about how it works!
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.
Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.
We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.
🧵👇
We’re happy to announce 2 releases today: - 🧠Brain2qwerty v1 is published at @NatureNeuro - 🚀 Brain2Qwerty v2 is now publicly released
Explore how we decode sentences from non-invasive brain recordings: https://facebookresearch.github.io/brain2qwerty/
Links: 📄v1 Nature Neuro: https://www.nature.com/articles/s41593-026-02303-2 📑v2 Meta preprint: https://facebookresearch.github.io/brain2qwerty/assets/brain2qwerty_v2.pdf 💻Code: https://github.com/facebookresearch/brain2qwerty 📊Data: https://huggingface.co/datasets/bcbl190626/SpanishBCBL 📝Blog: https://ai.meta.com/blog/brain2qwerty-brain-ai-human-communication/ 🧵Thread:
🧠⌨️ Decode language from brain activity without surgery. 🧠⌨️
Brain2Qwerty V1 is officially published in Nature Neuroscience. Today, we're releasing Brain2Qwerty V2. We achieve unprecedented performance for a non-invasive MEG setup.
Details below 🧵👇
i'm writing a thread about Meta's new brain-to-text model hope it's of interest :)
Something interesting about this project that not many people are paying attention to:
The researchers use autoresearch to improve their pipeline.
That's right: Meta is using AI to train an AI to read minds!
Meta researchers used LLMs to predict the text a person was typing just from non-invasive brain recording!
Literal mind reading!
"For our best participant, the model accurately decodes half of the sentences with one word error or less." → INSANE
Let's learn about how it works!
To help accelerate neuroscience breakthroughs, we're releasing the full training code for Brain2Qwerty v1 and v2, and our partner, @bcbl_, is releasing the v1 dataset.
Learn more and explore the artifacts here: https://go.meta.me/42ed9c
We trained Brain2Qwerty v2 on ~22,000 sentences from 9 volunteers, each recorded for 10 hours wearing an MEG device while typing.
By using end-to-end deep learning on raw brain signals from MEG devices and fine-tuning LLMs, the system effectively bridges the gap between noisy neural data and coherent language.
The results are promising: - Avg word accuracy of 61% across participants - 78% word accuracy and 50%+ of sentences decoded with ≤ 1 word error for the top-performing participant - Performance scales log-linearly with data volume
It's hard to read Meta working on something like this without tacking on "...and think how much better we'll be able to target ads like this!"
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.
Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.
We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.
🧵👇
Hooked the brain decoder up to my friend who posts thought leadership on Linkedin.
It registered nothing.
I complained to Meta support, and they said "oh, it's working."
SITUATION DETECTED: Meta has released Brain2Qwerty v2, a non-invasive brain-to-text decoder capable of real-time sentence decoding from raw brain signals.
Meta says the research could help millions of people with disorders that prevent communication.
Preparing detailed well-written threads like this takes a lot of time and research.
Too bad Twitter broke my thread when I posted it the first time so I had to repost and and now it's not being shown to anyone 😭
(Twitter doesn't like it when you do similar tweets)
Meta researchers used LLMs to predict the text a person was typing just from non-invasive brain recording!
Literal mind reading!
"For our best participant, the model accurately decodes half of the sentences with one word error or less." → INSANE
Let's learn about how it works!
@bcbl_ For clarification, Brain2Qwerty v1 was published earlier today in @NatureNeuro.
Vibe coding is gonna hit new levels of insanity
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.
Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.
We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.
🧵👇
@AIatMeta @Nature The last thing I'd want is Zuckerberg to have access to my brain even if I was disabled.
The Aligner uses a SigLIP contrastive loss to learn word-level alignment between the Encoder's MEG embeddings and the LLM's word embedding space.
Whenever the CTC path from the Encoder predicts a space, it chops up the continuous MEG into neural "words".
Because spaces are frequent (about 19% of characters) and robustly predicted, 86% of sentences have their word count estimated within ±1 word of the ground truth — which is what lets the LLM read the MEG as structured token input.
Let's dive deeper into each component.
The Encoder itself consists of two stages: a "BrainModule" that is a convolutional feature extractor, and a Conformer model.
This Conformer outputs a temporal MEG embedding. This MEG embedding is then mapped to discrete 28-character predictions with a linear layer. It is trained with a Connectionist Temporal Classification (CTC) objective which learns to align variable length MEG data to keystroke sequences without needing the exact timing.
Therefore, the Encoder outputs both an MEG embedding and character-level predictions.
Looks like there's a lot of interest, well here you go!
Meta researchers used LLMs to predict the text a person was typing just from non-invasive brain recording!
Literal mind reading!
"For our best participant, the model accurately decodes half of the sentences with one word error or less." → INSANE
Let's learn about how it works!
Finally, an LLM (Qwen3-4B) takes in both the MEG embeddings (chopped up at the word-level) and the characters decoded from the Encoder's CTC head. The LLM outputs the final decoded text.
The LLM is trained with a simple LoRA adapter. The model was trained in three phases: CTC-only, CTC+SigLIP, CTC+SigLIP+LLM cross entropy.
Separate LoRAs are trained for each subject before merging them into one LoRA.
The Aligner uses a SigLIP contrastive loss to learn word-level alignment between the Encoder's MEG embeddings and the LLM's word embedding space.
Whenever the CTC path from the Encoder predicts a space, it chops up the continuous MEG into neural "words".
Because spaces are frequent (about 19% of characters) and robustly predicted, 86% of sentences have their word count estimated within ±1 word of the ground truth — which is what lets the LLM read the MEG as structured token input.
