Pick your matchup.
Find a live or upcoming fixture across your favorite sports. One match. A whole new perspective.
Go beyond the score. Follow the action, explore the context behind each matchup, and approach your next prediction with a clearer view of the game.
SIMULATED SOURCE VOLUME
REAL EVIDENCE ATTACHED
Simulated volume · Real linked examples
Follow the numbers, team updates, and conversations that add context to a match.
Follow the action. Live scores and match clocks, updated as matches unfold.
Select a live match to get started. The prediction reports elsewhere on this page remain illustrative samples.
Public social posts inform the prediction. Source screenshots show the original posts, and the report explains how it reached its result. Here’s the experience we’re building.
Find a live or upcoming fixture across your favorite sports. One match. A whole new perspective.
The planned OpenCode pipeline researches public posts on X, Reddit, and Facebook, checking relevant claims against match context and other sources.
Get a prediction with screenshots of the posts behind it, source references, and an explanation of how each finding influenced the call.
The first thing I look for is where the information came from. Let me open the original post and make up my own mind about the evidence.
Before kickoff, I’m usually bouncing between Reddit, X, and a few Facebook groups. Having the useful conversations in one place would save me a lot of digging.
I always want to know why a prediction leans one way. What’s backing it up? What could change the call? That’s the part I’m interested in.
A few things to know about Picks Fillet.
The Match Center displays only in-progress matches, with real scores and match status from ESPN’s public scoreboards updated every 30 seconds while the tab is active. Scheduled and finished games are hidden. Coverage includes selected European football leagues, the NBA, ATP, and WTA. Provider updates can be delayed or unavailable. The hero report, prediction probabilities, and research examples remain illustrative; automated research is not connected.
Social media research is central to Picks Fillet. The planned AI system researches permitted public posts on X, Reddit, and Facebook to identify relevant discussion, sentiment, and potential updates, then checks those signals against match context and other sources. Source access depends on platform permissions and availability.
A predicted winner, estimated probabilities, screenshots of the source posts, and a clear explanation of how each finding informs the call. The planned live version will also include source permalinks, capture times, conflicting viewpoints, and uncertainty.
Yes. The Celtics–76ers sample includes real NBA, 76ers, ESPN, and Reddit sources for 10 November 2026, plus genuine screenshots of Reddit’s official embeds. Original links and capture times accompany the screenshots. The archive includes post-game records and discussion; these sources did not generate the illustrative probability.
No. Sports outcomes are uncertain, and predictions can be wrong. Picks Fillet is designed to explain the research so you can assess the evidence for yourself.
The planned OpenCode pipeline will coordinate source research, organize findings, and help produce a structured prediction report. This demo shows the intended experience; no live crawling or AI generation runs here.
Follow the conversation. See the source. Understand the prediction.
Predict live matches