How to build a market signal feed from finance YouTube

The problem it solves

Traders and researchers want to know which assets influential voices are discussing and whether sentiment is shifting, but the information is locked in long videos. By the time a clip spreads on social media, the move may have happened. A structured feed turns spoken commentary into data that can be scanned, searched and backtested.

What to include in your first version

• ■A curated list of analyst and commentator channels
• ■Ticker and asset extraction from every new video
• ■Bullish, bearish or neutral sentiment per mention
• ■A daily digest ranked by number of mentions and sentiment shift
• ■A JSON feed or API for quantitative users

How to build it, step by step

1. 01

   Curate channels

   Pick channels by category and influence. Resolve each one once to store its channel ID.

2. 02

   Poll hourly

   Resolve each channel with a small limit and transcribe only new video IDs with /transcribe.

3. 03

   Extract structured signals

   Ask an LLM to return JSON listing each asset mentioned, the claim made, the sentiment and the timestamp.

4. 04

   Aggregate and publish

   Store signals in a database, rank them daily, and publish a newsletter, dashboard or API feed.

Starter code

This idea is built on the "Recipe 2: Watch channels and alert on new uploads" pattern. Swap in your own channels, prompts and storage.

import requests

API = "https://www.youtubetranscript.dev/api/v2"
HEADERS = {"Authorization": "Bearer yt_sk_live_YOUR_KEY"}

WATCHLIST = ["@somechannel", "@anotherchannel"]
KEYWORDS = ["your brand", "competitor"]
seen = set()  # persist this in a database in production

for handle in WATCHLIST:
    latest = requests.post(f"{API}/channels/resolve", headers=HEADERS,
                           json={"handle": handle, "limit": 10}).json()
    for video in latest["items"]:
        if video["video_id"] in seen:
            continue
        seen.add(video["video_id"])
        t = requests.post(f"{API}/transcribe", headers=HEADERS,
                          json={"video": video["video_id"]}).json()
        text = t.get("data", {}).get("transcript", {}).get("text", "").lower()
        hits = [k for k in KEYWORDS if k in text]
        if hits:
            print(f"ALERT {video['title']}: {hits}")

How it makes money

Run it as a paid newsletter for retail traders, a dashboard subscription for research teams, or an API feed priced by coverage. Historical data becomes more valuable over time for backtesting.

Mistakes to avoid

• ✕Presenting signals as investment advice: label the feed as information and include clear disclaimers.
• ✕Confusing tickers with common words: validate symbols against a ticker list and the surrounding context.
• ✕Storing only summaries: keep raw transcripts so you can re-extract signals as your prompts improve.

Questions about this idea