NeuralSEO

Fresh

Published: March 30, 2026 | Author: Metehan Yesilyurt

NeuralSEO applies neuroscience to SEO by using Meta AI's TRIBE v2 brain scanner model to predict how human brains process search results. Instead of guessing what catches attention, NeuralSEO measures it using brain signal predictions.

Access: huggingface.co/spaces/metehan777/neuralseo

Source

Based on Metehan's original article and the TRIBE v2 model card.

What is TRIBE v2

TRIBE v2 stands for TRImodal Brain Encoder v2. It is a foundation model from Meta AI's FAIR lab, trained on fMRI recordings — actual brain scans — collected while 700+ volunteers watched videos, listened to audio, and read text. It is a multilanguage model.

The model learned to predict how the human cortex responds to any input across roughly 20,000 cortical vertices. Feed it a sentence, and it tells you which brain regions activate, how strongly, and for how long.

Architecture

TRIBE v2 combines three state-of-the-art feature extractors into a unified Transformer:

Component	Role
LLaMA 3.2-3B	Text encoder
V-JEPA2	Video encoder
Wav2Vec-BERT 2.0	Audio encoder

The model maps multimodal representations onto the cortical surface with predictions for the fsaverage5 cortical mesh (~20k vertices).

Quick Start (Python)

from tribev2 import TribeModel

model = TribeModel.from_pretrained("facebook/tribev2", cache_folder="./cache")

df = model.get_events_dataframe(video_path="path/to/video.mp4")
preds, segments = model.predict(events=df)
print(preds.shape)  # (n_timesteps, n_vertices)

Input flexibility: pass text_path, audio_path, or video_path. Text is automatically converted to speech with word-level timings.

Installation

# Inference only
pip install -e .

# With brain visualization (PyVista, Nilearn)
pip install -e ".[plotting]"

# With training dependencies (PyTorch Lightning, W&B)
pip install -e ".[training]"

What NeuralSEO Does

Metehan built three core analysis tools on top of TRIBE v2.

1. Neural Screenshot Analyzer

Upload a screenshot of a Google SERP, ChatGPT response, Perplexity answer, or Google AI Mode result. The tool:

1. Splits the screenshot into layout regions (title, snippet, sidebar, etc.)
2. Crops each region and runs it through TRIBE v2's visual inference pipeline
3. Scores each element by neural attention activation
4. Draws live scored overlays directly on the image

TIP

This is the closest thing to eye-tracking without actual eye-tracking hardware.

2. Intro Paragraph Analyzer

Paste your opening paragraph (auto-trimmed to 600 characters). TRIBE v2 scores it across four neural dimensions:

Dimension	What It Measures
Hook Strength	Does the opening trigger frontal attention networks?
Engagement	Global neural activation level
Salience	Does it stand out from noise?
Retention	Will the reader's brain encode this into memory?

You get a 0-100 neural score, a radar chart breakdown, and optional Gemini-powered rewrite recommendations.

3. Neural CTR Predictor

Enter a keyword. Gemini generates 10-20 dynamic title tag variants. Each title runs through TRIBE v2 individually, scored by frontal attention network activation and salience response. You get a ranked list of predicted organic CTR before you publish — without A/B testing.

Brain Signal to SEO Mapping

Brain Signal	SEO Metric Proxy	Interpretation
Language comprehension activation	Readability clarity	Smooth language processing = lower bounce
Frontal attention networks	Bounce/retention risk	High frontal activation = engaged reader
Activation entropy	E-E-A-T proxy	Expert content produces different activation patterns than thin content
Salience network	Title attention demand	High salience = higher CTR
Default Mode Network (inverse)	Mind-wandering/bounce risk	DMN suppression = focused attention

WARNING

These are neurological proxies — directional signals based on how the human brain processes content. Google's ranking algorithm doesn't use fMRI data directly.

Technical Architecture

Stack Details

Component	Technology
Model	facebook/tribev2 — Meta's trimodal brain encoder
Text pipeline	Input → TTS audio → word-level timestamps via faster-whisper → TRIBE v2 fMRI prediction
Visual pipeline	Image → short MP4 via moviepy → TRIBE v2 visual inference
Title generation	Google Gemini Flash generates dynamic variants, TRIBE v2 scores them
Frontend	Gradio with custom dark theme, procedural Three.js brain visualization
Brain viewer	Procedural mesh with 5 cortical regions that light up based on analysis scores
Hosting	Hugging Face Spaces with GPU allocation

Text Pipeline Detail

TRIBE v2 was trained on multimodal data, so even for text analysis, the input goes through a TTS step to generate audio, which is then transcribed with word-level timestamps. This gives the model the temporal dynamics it needs to predict brain activation patterns over time.

TRIBE v2 Project Structure

tribev2/
├── main.py              # Experiment pipeline
├── model.py             # FmriEncoder: Transformer-based multimodal→fMRI
├── pl_module.py         # PyTorch Lightning training module
├── demo_utils.py        # TribeModel and inference helpers
├── eventstransforms.py  # Event transforms (word extraction, chunking)
├── utils.py             # Multi-study loading and splitting
├── utils_fmri.py        # Surface projection (MNI/fsaverage) & ROI analysis
├── grids/
│   ├── defaults.py      # Full default configuration
│   └── test_run.py      # Quick local test entry point
├── plotting/            # Brain visualization (PyVista & Nilearn)
└── studies/             # Dataset definitions

Limitations

• Neural scores are directional signals, not ground truth ranking guarantees
• GPU quotas are real — on Hugging Face's free tier, large batches can timeout. Use smaller inputs when possible
• First request is slow — the TRIBE v2 model weighs ~6 GB and loads on the first inference call
• Non-commercial use only — TRIBE v2 is licensed CC BY-NC 4.0
• Predictions are based on averaged brain responses from 700+ study participants — individual variation exists
• Works best for relative comparisons (variant A vs. variant B) rather than absolute scores

Resources

Resource	Link
NeuralSEO Tool	huggingface.co/spaces/metehan777/neuralseo
TRIBE v2 Model	huggingface.co/facebook/tribev2
Research Paper	Meta AI Research Publications
Interactive Demo	Meta AI Demo
Colab Notebook	Google Colab
Original Article	metehan.ai blog post

Citation

@article{dAscoli2026TribeV2,
  title={A foundation model of vision, audition, and language
         for in-silico neuroscience},
  author={d'Ascoli, Stéphane and Rapin, Jérémie and
          Benchetrit, Yohann and Brookes, Teon and
          Begany, Katelyn and Raugel, Joséphine and
          Banville, Hubert and King, Jean-Rémi},
  year={2026}
}