Nadelio, AI Visibility Intelligence

A brand visibility score you can actually show your board, not noise.

Try Nadelio

When a buyer searches "best tool for X", who shows up, you or your competitors? And when they ask the same question to an AI assistant, who gets recommended?

Nadelio answers both with live data. You type a single brand name; the system identifies your competitive set, runs live Google searches, checks AI engine recommendations, and renders a dual heatmap of SERP ranks against AI mentions, side by side. This is the measurable side of GEO / AEO (Generative Engine Optimization).

The difference: most tools hand you a single number that is really a noisy sample of one on a non deterministic model. Nadelio measures across multiple runs and reports a bounded score (a value with its confidence), so you know whether a move is real or just noise.

---

How it works

A 4-step pipeline:

  "Notion"  ->  Claude  ->  competitors + buyer-intent queries
                                  |
                                  v
                          Bright Data SERP API  ->  live Google results
                                  |
                                  v
                          Claude AI check  ->  AI engine recommendations
                                  |
                                  v
                    dual heatmap, share of voice, AI engine leader

1. Competitive intelligence. A single brand name goes to Claude, which infers the sector, real competitors, and the buyer-intent queries a customer would search.
2. Live SERP data. Each query runs against Google via Bright Data. The app finds where every brand ranks.
3. AI engine check. Claude checks how an AI assistant would answer the same buyer questions, revealing which brands get recommended in AI-powered search.
4. Dual heatmap. Green cells for SERP ranks, purple cells for AI mention order, plus share of voice, coverage, average rank, and clickable evidence.

Key features

• Bounded score. A visibility score reported with its confidence across multiple runs, not a single noisy number.
• Dual visibility. SERP and AI recommendations in one view. See the gap between traditional and AI-powered search.
• Zero configuration. Type one brand name, get a full competitive analysis in about 10 seconds.
• Persistent cache. Analyses are cached for instant re-access.
• Pro plans coming soon. Ongoing monitoring, alerts, and expanded analysis.

Run locally

git clone https://github.com/arochab/nadelio.git
cd nadelio
pip install -r requirements.txt

export BRIGHTDATA_API_KEY="your_bright_data_key"
export ANTHROPIC_API_KEY="your_anthropic_key"
python app.py
# open http://localhost:5000

Without API keys, the app runs in sample mode with precomputed data.

Also in this repo

Command-line tools the web app builds on:

# Single Google query to structured JSON / CSV
python serp_scraper.py "your keyword" --csv

# Multi-brand visibility report from the terminal
python ai_visibility.py --config visibility_config.example.json

License

MIT · Built by Adam Chabbi