Market Intel
The AI Tool That Searches Twitter for Market Intelligence
March 17, 2026
Every experienced crypto trader knows the same thing: price moves after narrative, and narrative forms on X/Twitter. The FTX collapse didn't start on Bloomberg. It started with a CoinDesk article that got amplified by CT accounts who spotted the Alameda balance sheet anomaly. The traders who were monitoring X sentiment in real-time exited before the cascade. Everyone else watched from underwater.
The problem is that monitoring X/Twitter manually doesn't scale. You can follow 200 accounts and still miss the one thread from a researcher with 400 followers who identified a critical vulnerability. Lists help. TweetDeck helps. But none of it synthesizes the signal—you still have to read hundreds of tweets and decide what matters.
X/Twitter as a Trading Signal
In traditional markets, Bloomberg terminals aggregate news and data into actionable signals. Crypto doesn't have that. The closest equivalent is X/Twitter, where project teams announce updates, whales discuss positions, security researchers disclose vulnerabilities, and communities signal confidence or panic—all in real-time, all public.
But X is noisy. For every legitimate alpha leak, there are 50 engagement-farming threads, paid promotions disguised as analysis, and bot networks amplifying low-quality projects. The signal-to-noise ratio is terrible if you're reading it raw. You need a filter—something that can separate the security researcher's thread from the influencer's paid shill, and cross-reference both against on-chain data and web sources.
How Probe Searches X/Twitter
Probe's research engine searches both the open web and X/Twitter simultaneously through xAI's multi-agent system. When you run a query like “What are whales doing with $ETH right now?”, some agents scan blockchain analytics sites and news outlets while others pull real-time tweets from whale-tracking accounts, on-chain analysts, and relevant discussions.
The key difference from scrolling your timeline: Probe synthesizes across sources. It doesn't just show you tweets—it tells you that 3 whale-tracking accounts flagged a 15,000 ETH transfer to Binance, DeFiLlama shows a $40M TVL drop in the last 6 hours, and CT sentiment has shifted from bullish to cautious based on 47 unique accounts. That's the kind of synthesis that takes a human 2 hours and an AI system 60 seconds.
Every claim is cited with a source link. You can click through to the original tweet or article to verify. The report isn't asking you to trust the AI—it's giving you a structured summary with receipts.
Real-World Use Cases
Scheduled Monitoring
One-off queries are useful, but the real edge comes from continuous monitoring. Probe supports scheduled research—you can set up recurring queries like “Daily CT sentiment report on $ETH, $SOL, $BTC” or “Twice-weekly whale activity summary for my portfolio tokens.” Each run produces a fresh report searching both web and X/Twitter, and you get notified by email when it's ready.
This turns Probe into an always-on research assistant. Instead of checking X/Twitter yourself every morning, you wake up to a synthesized intelligence brief with cited sources, sentiment analysis, and flagged changes from the previous report.
More from the blog
New accounts get $5 in free credits. Run your first X/Twitter intelligence query in under 60 seconds.