Skip to main content
The main Research view is one feed across Substack + X + YouTube. The controls let you slice it.

​
Sort

Three sort orders, switchable from the top:

​
Filters

​
By source

Toggle Substack / X / YouTube on or off independently. Useful when you want only long-form (Substack), only quick takes (X), or video. A fourth segment — the 💬 chat bubble — narrows to Substack chat threads: the subscriber chats of paid publications (AMAs, live trade talk), each ingested as a transcript. Chat items carry a CHAT badge in the feed, and their article page includes the full collapsible transcript with a message count.

​
By ticker

Type a ticker to see only content mentioning it. Multi-ticker is supported.

​
By author / publication

Narrow to a specific writer or publication.

​
By confidence tier

Filter to High / Medium / Low confidence. → What confidence tiers mean

​
By date range

Default is last 7 days. Expand to 30, 90, or pick a custom range.

​
What each card shows

Per item in the feed:
  • Title (or first 240 chars for X posts that don’t have titles)
  • Author + publication — clickable to filter to that author
  • Time ago (“3h ago,” “2 days ago”)
  • Confidence badge (color-coded by tier)
  • Tickers tagged — primary tickers (the focus) vs mentioned (passing references)
  • Source badge — Substack / X / YouTube
  • Special tags — THREAD, RT (retweet), QT (quote-tweet), PAID, etc.
Click the title to read the full piece in a side panel without leaving the page. The search box runs semantic search across the corpus — type a phrase like “uranium supply” or “Fed pivot” and get the top matches by meaning, not just keyword. Behind the scenes this hits the pgvector embeddings. Same engine the Claude Desktop MCP uses.

​
How most people use it

  • Morning — sort by Newest, filter by date = today, scan the top 20
  • Researching a name — filter by ticker, sort by Confidence, last 90 days
  • Following a theme — search the phrase, sort by Newest