Find the conversations worth joining.

A local, personalized filter for concise X conversations with real reply momentum, a clear opening, and room for a practical contribution.

Scheduling next scan…

No completed scans yet

Official X APIBroad RecallLocal FitReply MomentumQuality GatesTop Conversations

Public post data is retrieved through the official X API. No scraping, automated engagement, model training, or external AI processing. This independent project is not affiliated with or endorsed by X.

Best current conversations

Reading the latest signals…

01CollectOfficial X API
Searches recent public, English, original posts across AI tools, software, Git, SaaS, building, and product-platform discussions. The query favors direct questions, debatable positions, switching decisions, friction, and fast-changing claims.
Why: Retrieval uses only the documented X API—never scraping or browser automation—and asks for the minimum public fields required to rank and render a compliant post.
02MinimizeValidation
Rejects replies, reposts, quote posts, article posts, paid nullcasts, link-heavy posts, malformed records, missing display attribution, Ali’s own posts, and post IDs covered by a removal request.
Why: The product keeps only data needed for this narrow owner tool and never accesses protected posts, Direct Messages, private metrics, location, or sensitive inferred attributes.
03UnderstandLocal ranker
Scores personal topic fit, whether the post creates a real conversational opening, and brevity. The profile favors concise questions, grounded disagreements, product-switching decisions, builder trade-offs, developer workflows, and social-product debates.
Why: All analysis runs locally. X Content is not sent to an external AI provider and is never used to train or fine-tune a model.
04MeasurePublic metrics
Measures reply momentum, replies relative to public views, and reach. Likes, reposts, quotes, and bookmarks remain visible context but cannot substitute for an active conversation.
Why: The target is a useful exchange, so replies and their pace matter more than passive popularity or follower authority.
05FilterQuality gates
Requires topic fit, a strong opening, at least three public replies, enough overall opportunity score, and either concise text or exceptional momentum. Only the six best qualifying conversations are returned; Skip results are withheld.
Why: Returning nothing is better than filling the page with long, quiet, weakly matched posts. Long form earns an exception only when reply activity and reach are genuinely exceptional.
06DisplayX attribution
Renders the full unmodified post text with linked entities, author avatar, display name, @username, timestamp, official X mark, public metrics, and a titled X permalink.
Why: The custom rendering follows X’s display requirements. The optional Suggest reply control only copies a local writing brief; it never calls an AI provider or posts, likes, reposts, follows, or shares anything on X.
07ExpireRailway volume
Caches only the latest qualifying snapshot, scan counters, and a content-free last-scan timestamp. An independent timer removes X Content within 24 hours at the absolute latest, while that timestamp preserves the 10-day scan cadence after expiry.
Why: Short retention minimizes stored X Content and bounds how long a deletion, protection change, suspension, withholding, or edit could remain visible between official API refreshes.
08RemoveOwner purge
A protected backend action immediately removes a requested post from the snapshot and stores only a one-way hash so the same post cannot reappear in later scans.
Why: Account-owner and X removal requests can be honored promptly without retaining the removed post text, author profile, or recoverable identifier.
09ObserveBetter Stack
Records only scan success, safe counts, duration, and failure category. Logs never contain post text, post or author identifiers, credentials, owner tokens, or visitor data.
Why: The system remains diagnosable without redistributing or leaking X Content through monitoring.

Opportunity Score weights

Personal fit: 22%
Conversation opening: 22%
Reply momentum: 22%
Reply density: 14%
Brevity: 12%
View reach: 8%

How the radar works

Best current conversations