How to build a YouTube brand monitoring tool

The problem it solves

Influential opinions about products are increasingly spoken in reviews, podcasts and commentary videos, often without the brand name in the title. PR teams find out days later, if at all. They need alerts when a brand is mentioned, the context of what was said, and a way to measure share of voice against competitors.

What to include in your first version

• ■A watchlist of channels and podcasts per client
• ■Keyword and alias matching for brands, products and people
• ■AI classification of each mention as positive, negative or neutral
• ■Slack and email alerts with the quote and a timestamped link
• ■Weekly share-of-voice reports versus competitors

How to build it, step by step

1. 01

   Build the watchlist

   Store the channels to monitor for each client. Resolve each handle once with /channels/resolve to get the channel ID.

2. 02

   Poll for new uploads

   On a schedule, resolve each channel with a small limit and compare the video IDs against those you've already processed.

3. 03

   Transcribe only new videos

   Send new IDs to /transcribe or /batch. You spend credits only on videos you haven't seen.

4. 04

   Detect and classify mentions

   Search transcripts for brand names and aliases, then pass the surrounding paragraph to an LLM to confirm relevance and classify sentiment.

5. 05

   Alert and report

   Send each confirmed mention to Slack or email with the quote and a timestamped link, and roll mentions up into weekly reports.

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

Sell monthly subscriptions priced by the number of brands tracked and channels monitored. Agencies managing several clients are the strongest buyers; a weekly report alone can justify the plan.

Mistakes to avoid

• ✕Sending raw keyword matches as alerts: names like "Apple" or "Notion" need an LLM relevance check to avoid noise.
• ✕Monitoring too few channels: the value is coverage, so let clients add channels easily.
• ✕Missing mentions in videos without captions: enable AI speech recognition for important channels.

Questions about this idea