FinTwit Scanner Guide
AI-powered Twitter intelligence for smarter trading decisions
What is the FinTwit Scanner?
The FinTwit Scanner monitors 100+ curated financial Twitter/X accounts and uses AI to extract actionable stock picks from their tweets. Every tweet is analyzed individually for stock symbols, sentiment (bullish, bearish, or neutral), and a confidence score — giving you a real-time view of what top traders are thinking.
AI-Powered Analysis
Gemini 2.0 Flash analyzes every tweet, extracting stock symbols, sentiment classification, confidence scores (0–100%), and reasoning.
100+ Expert Handles
Curated list of macro traders, chart technicians, sector specialists, value investors, and momentum traders.
Signal Integration
Use "Twitter Picks" as a focus in The Agent — the AI synthesizer blends Twitter sentiment with technical analysis for entry/stop/target levels.
Real-Time Heatmap
Interactive treemap showing mention volume by sector and stock, color-coded by sentiment. Click sectors to drill down.
How It Works
The scanner runs a 5-step pipeline that turns raw tweets into actionable trading intelligence:
1Monitor FinTwit DATA
Automated scanning of 100+ curated financial Twitter/X accounts on configurable schedules (default: every 24 hours). New tweets are fetched via the Twitter/X API.
2AI Tweet Analysis LLM
Each tweet is analyzed individually by Gemini 2.0 Flash (primary) or OpenAI (fallback). The AI extracts: stock symbols mentioned ($TICKER, #TICKER, or contextual), sentiment (bullish/bearish/hold), confidence score (0–100%), and reasoning for the classification.
3Enrich Metadata CODE
Every extracted stock symbol is automatically enriched with sector, industry, and company name from our market data feed. This powers the heatmap grouping and sector/industry filtering.
4Store & Aggregate DATA
Recommendations are stored with full provenance: original tweet URL, tweet text, analysis timestamp, LLM reasoning, and which model analyzed it. An in-memory cache powers the heatmap for fast rendering.
5Deliver Insights CODE
View results via the interactive heatmap, filterable recommendation feed, configurable email alerts, shareable snapshot cards, or as "Twitter Picks" in The Agent's trading signal.
Every recommendation links back to the original tweet. You see the AI's reasoning, which model analyzed it, and the confidence score — so you can always verify the source.
Key Features
Confidence Scoring
Each recommendation gets a 0–100% confidence score. Explicit $TICKER mentions score highest; contextual mentions score lower. Filter out noise by setting a minimum threshold.
Filtering & Search
Filter recommendations by handle, ticker, sector, industry, or date range. Sort by date or ticker to find exactly what you need.
Manual Tweet Analysis
Paste any tweet URL to get instant AI analysis. Works with any Twitter handle, not just tracked ones — useful for evaluating tweets you encounter in the wild.
Stock Metadata
Every mentioned stock is automatically tagged with sector, industry, and company name from our market data feed. This powers the heatmap grouping and makes sector-level filtering possible.
LLM Reasoning
Every recommendation includes the AI's explanation. Understand why a tweet was classified as bullish or bearish, and what specific language triggered the classification.
Multi-Stock Tweets
A single tweet can mention multiple stocks with different sentiments. The AI extracts each stock-sentiment pair as a separate recommendation with its own confidence score.
Using Twitter Picks in Your Signal
The FinTwit Scanner feeds directly into The Agent's trading signal. When you select "Twitter Picks" as your focus, the AI synthesizer blends Twitter sentiment alongside technical analysis to generate entry, stop-loss, and target levels.
- Navigate to The Agent page
- In the focus selector, click the Twitter Picks button (the X/Twitter icon)
- The agent will prioritize stocks recommended by tracked FinTwit accounts
- Twitter sentiment data is blended with technical analysis for entry/stop/target levels
- Configure the lookback window (1–30 days) in Agent Settings to control how far back to scan for Twitter picks
Combine Twitter Picks with your custom watchlist. When both are active, you get recommendations that are either on your watchlist OR trending on FinTwit — widening your opportunity set without losing focus.
The Sentiment Heatmap
The heatmap is an interactive treemap visualization that gives you an instant view of what FinTwit is talking about and how they feel about it.
How to Read the Heatmap
- Tile size represents mention volume — more mentions from tracked handles means a bigger tile
- Green tiles indicate net bullish sentiment from FinTwit experts
- Red tiles indicate net bearish sentiment
- Gray tiles indicate mixed or neutral sentiment
- Click any sector to drill down and see individual stocks within that sector
- Configure the lookback period (1–365 days) and minimum confidence threshold to tune what appears
Large, bright green tiles indicate stocks with many bullish mentions from multiple experts — these are consensus picks. Large red tiles indicate bearish consensus. Small tiles with mixed colors represent stocks with divided sentiment, which may warrant closer investigation.
Email Alerts
Stay informed without checking the scanner constantly. Configure email subscriptions to get notified about new recommendations that match your criteria.
Subscription Options
- Frequency — Choose between immediate (as found), daily digest, or weekly summary
- Handle filter — Subscribe to specific handles you trust, or receive alerts from all tracked handles
- Stock filter — Subscribe to specific stocks you're watching, or receive all recommendations
- Minimum confidence — Set a threshold (default 30%) to only receive alerts above a certain confidence level
Create, update, or delete subscriptions from the Twitter Scanner page.
Email alerts require a Pro or higher subscription tier. The scanner page itself also requires Pro+ access.
Sharing Your Insights
Generate shareable snapshot cards from the Twitter Scanner page to share your FinTwit intelligence on social media.
How Snapshots Work
- Click the Share button on the Twitter Scanner page
- The system generates a snapshot card with your top recommendations and sentiment breakdown
- Each snapshot gets a unique public URL that anyone can view without logging in
- Share directly to Twitter/X, LinkedIn, or copy the link
Shared snapshots include a call-to-action for viewers to start their own free trial. Sharing regularly helps build your trading community and gives followers a preview of the platform's capabilities.
Requesting New Handles
Don't see your favorite FinTwit expert in the tracked list? You can request new handles to be added.
How to Request a Handle
- From the Twitter Scanner page, click Request Handle
- Enter the Twitter/X handle (without the @ symbol)
- Add optional notes about why this handle is valuable (what type of analysis they provide)
- An admin reviews the request and either approves or rejects it
- Approved handles are added to the monitoring list and begin scanning within 24 hours
Handle Categories Currently Monitored
- Macro traders — Market commentators and macro-level analysis
- Chart technicians — Technical analysis and pattern recognition
- Momentum traders — Breakout and momentum-based picks
- Value investors — Fundamental analysis and valuation
- Sector specialists — Deep expertise in specific sectors (biotech, tech, energy)
- Small/micro-cap specialists — Under-the-radar stock picks
When requesting a handle, include a note about what type of analysis they provide. This helps the admin team evaluate whether the handle adds unique value to the existing coverage.
FAQ
How often are handles scanned?
The default scan frequency is every 24 hours, but admins can configure different frequencies per handle. High-value handles that post frequently can be scanned more often.
Which AI model analyzes tweets?
Gemini 2.0 Flash is the primary model for tweet analysis, with OpenAI as the automatic fallback. The platform's prompt admin system manages the dynamic prompts and model selection, so the analysis keeps improving over time.
Can I submit a tweet for analysis manually?
Yes. Admins can paste any tweet URL into the scanner to get instant AI analysis, even from handles that aren't in the tracked list. This is useful for quick one-off evaluations of tweets you encounter.
What does the confidence score mean?
The confidence score (0–100%) reflects how strongly the AI believes the tweet contains a genuine stock recommendation. Scores above 70% indicate strong conviction. Scores between 30–70% suggest a moderate signal. Below 30% typically means a casual mention rather than a deliberate recommendation.
Why are some stocks missing sector/industry data?
Metadata is fetched from our market data feed after stock symbols are extracted. Some tickers — especially ETFs, very new IPOs, or penny stocks — may not have complete metadata available. These stocks still appear in the recommendations but won't be grouped by sector in the heatmap.
Do I need a subscription to use the Twitter Scanner?
The scanner page and heatmap require a Pro or higher subscription tier. This guide is public. Using Twitter Picks as a focus in The Agent also requires Pro+ access.