MCP

AI-agent format

Triage earnings season across your watchlist

Import the calls and interviews that matter, batch key-quotes and summaries, and walk into the week with a stack of one-pagers.

AnalystFinance, Investing & Due Diligencelibrary_import_youtube_videorun_youtube_toollibrary_list
The use case

The problem

Earnings season is a bandwidth problem wearing a calendar costume. Twenty names on the watchlist, each with a management interview here, a post-print CNBC hit there, a fireside at a broker conference — sixty-plus hours of executive commentary landing inside two weeks, all of it potentially material, none of it watchable at that volume.

So triage happens by proxy: you watch the three names you're closest to and read headlines for the rest. Which works until the quarter where the thing that mattered was said in minute forty of an interview for a name you skimmed. The tape was public. You just didn't have a way to process all of it.

The mechanics

How veridive MCP helps

library_import_youtube_video captures the tape as it drops: each interview and appearance imported into your veridive library, transcript included, tagged by ticker and quarter. library_list keeps the stack organized — filter by tag, see what's in, what's processed, what's pending, free.

run_youtube_tool is the processing line: summarizer turns each hour into a structured brief, key-quotes extracts the sentences that move theses — guidance color, demand language, the CFO's word choice on margins. Batch it across the stack and every name on the watchlist gets the same treatment: a one-pager with quotes, each traceable back to its source video. Coverage stops being a function of your favorite names.

Step by step

How to run it

  1. 1
    Connect veridive MCP

    One-time setup: add the veridive MCP server to Claude, ChatGPT, Cursor or any MCP client and generate your API key. Takes about two minutes.

  2. 2
    Stack the tape as it drops

    Import each management appearance the day it publishes, tagged by ticker and quarter so the stack stays sorted.

    Import this interview into my veridive library, tagged '<TICKER>' and '<quarter>': <video URL> — use veridive's library_import_youtube_video.
  3. 3
    Run the processing line

    Batch summaries and key quotes across the imported stack. Each name gets a one-pager regardless of how close you are to it.

    For each video tagged '<quarter>' in my library: a structured summary plus the key quotes on guidance, demand and margins.
  4. 4
    Read, rank, go deep where it's earned

    The one-pagers tell you which three names deserve your actual hours this quarter — triage by evidence instead of familiarity.

    List everything in my library tagged '<quarter>' and flag which summaries contain surprises versus prior-quarter language.
Before vs after

Problems it solves

  • Sixty hours of earnings-season commentary lands in two weeks; human coverage caps at three names.
  • Headline coverage strips the language texture — hedges, word swaps, tone — where the real signal lives.
  • Notes from last quarter's interviews are unfindable exactly when this quarter needs the comparison.
  • Coverage bias: the names you know best get deeper reads, orphaning the rest of the list.
The payoff

Why it's valuable

  • Every watchlist name processed identically — triage driven by content, not familiarity.
  • Key quotes arrive attributed to their video, ready for the morning note without a fact-check scramble.
  • The library persists: next quarter's "what changed in their language?" runs against this quarter's stack.
  • Hours of tape per name compress to minutes of reading, with the source one click away.
Copy & paste

Example prompts

Import this earnings interview, tagged '<TICKER>' '<quarter>': <URL> — use veridive's library_import_youtube_video.
Summaries + key quotes on guidance and margins for every video tagged '<quarter>'.
List my '<quarter>' imports and flag which contain language changes versus last quarter.
Questions

FAQ

Does this work on the earnings calls themselves?

It works on whatever is on YouTube — which increasingly includes call replays, and always includes the management media rounds (network hits, broker firesides) where executives ad-lib beyond the script. That ad-lib layer is where triage finds surprises.

Why import instead of just searching each interview?

Volume and recurrence. The stack gets processed in batches, stays tagged by quarter, and becomes the baseline the next quarter's language gets compared against. Search is for discovery; the library is for coverage.

Is this investment advice?

No — it's coverage infrastructure. It compresses public commentary into readable, cited briefs; the judgment about what any of it means for a position stays entirely yours.

Keep going

Related use cases

Run this in your own assistant

Connect the veridive MCP server to Claude, ChatGPT or Cursor and paste the prompts above — setup takes about two minutes.