Inspiration
We are big fans of Community Notes on X (Twitter) because they fact check by adding context to posts leveraging the power of people. We are also fans of Prediction Markets as a source of truth and search engine since participants have skin in the game. We wanted to combine them.
What if you could see not just what someone tweets, but what they actually believe enough to invest in?
What it does
Market Notes transforms X (Twitter) into a prediction market-powered platform where claims meet consequences and opinions are backed by real investment data. Here's how it works:
On every tweet, a Polymarket icon appears alongside the standard like, retweet, and reply buttons. When clicked, our AI instantly analyzes the tweet's content and surfaces the most relevant prediction markets related to the tweet. Users can see what the current odds are for each market right and place a trade without leaving X. Our interface shows live pricing, calculates potential profits, and executes real-time orders through Polymarket's trading infrastructure.
Profile-level position tracking: On any user's X profile, you can see their actual Polymarket positions - both open and closed trades (if they have linked their X and Polymarket accounts). This provides unprecedented context: when @CryptoInfluencer tweets bullish takes on Bitcoin, you can see they have $5,000 riding on BTC markets, adding context to their tweets.
With Market Notes, you can fact check outdated, incorrect, or biased tweets by following the money. See a tweet from someone you trust about news that hasn't hit the market yet? Make the trade.
How we built it
We created a Chrome extension to add native feeling Polymarket widgets to X. User presses the Polymarket icon: Tweet gets sent through Cohere Command R + model to perform sentiment analysis to generate a concise query that we feed into the Polymarket API to fetch relevant events and markets. We take the list of JSON outputs and feed it back into the model to compare against the original query, and take the top 5 most relevant markets to the tweet.
Integrated with Polymarket's Central Limit Order Book (CLOB) API using py-clob-client and Magic wallet authentication. This enables real trading capabilities including live position tracking, order execution, and portfolio management. The system handles complex market data formats (singular and multi-market events), price validation, and trade execution while maintaining security through protected API keys.
We use various endpoints from Polymarket to grab the open and closed positions to display for the user on their profile.
Challenges we ran into
• Ensuring consistent data structures being passed around.
• Burning through tokens too fast.
• Getting Polymarket setup (may or may not have required a VPN).
• Market Relevance Scoring: Figuring out a reliable way to get consistently not bad markets is hard.
Accomplishments that we're proud of
• End-to-End AI Integration: Successfully built a complete pipeline from tweet text to relevant prediction markets using advanced NLP with Cohere AI, including sentiment analysis, market search, and relevance ranking.
• Live Trading Capabilities: Integrated real trading functionality with Polymarket's API using Magic wallet authentication, allowing users to execute actual trades directly from X.
• Adaptive UI System: Some UI parts actually look like components that would be used on X.
What we learned
• AI Model Behavior: We learned that AI relevance scoring works best with specific prompts and temperature settings.
• Prediction Market Dynamics: Working with live market data taught us about prediction market mechanics, price validation, and the importance of handling edge cases like markets with <1% probabilities.
• User Experience Design: We learned that complex financial interfaces need to be simplified for social media contexts. Our iterative design process showed the importance of clear visual hierarchies and error states.
What's next for market-note
• Authenticating accounts (right now you'd have to clone and add your own keys to run it). This way users can actually link X and Polymarket
• Selling positions
• GRAPHS!!!!!!!!
• Market Creation Tools: Allow users to create new prediction markets directly from tweets when relevant markets don't exist, democratizing the market creation process (right now markets are suggested and created on the Polymarket Discord server)
Built With
- api
- chrome
- clob
- cohere
- css
- dom
- flask
- git
- html
- http
- javascript
- json
- natural-language-processing
- polygon
- py-clob-client
- python



Log in or sign up for Devpost to join the conversation.