How to build an influencer vetting tool for YouTube sponsorships

The problem it solves

Influencer teams vet creators by watching a handful of recent videos. They miss a competitor sponsorship from six months ago, a controversial remark, or a habit of mocking sponsors on camera. Discovering this after a campaign launches is expensive. Reviewing hundreds of hours manually is impossible.

What to include in your first version

• ■Bulk analysis of a creator's recent videos
• ■Detection of past sponsors and competitor brands
• ■Content-safety flags for profanity, sensitive topics and controversy
• ■Sentiment toward sponsors and product categories
• ■A shareable report with quotes and timestamped evidence

How to build it, step by step

1. 01

   Pull the back catalogue

   Call /channels/resolve with the creator's handle and a limit of 100 to 500 to get recent video IDs.

2. 02

   Transcribe in one batch

   Send the IDs to /batch. For videos without captions, ASR options such as content moderation and sentiment analysis add structured safety signals.

3. 03

   Extract sponsors and risks

   Run an LLM over each transcript to list sponsors mentioned, competitor brands and statements that match the brand's risk categories.

4. 04

   Score and report

   Combine findings into a score per category and generate a report where every flag links to the quote and timestamp.

Starter code

This idea is built on the "Recipe 3: Bulk-load a playlist into a knowledge base" 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"}

playlist = requests.post(f"{API}/playlists/resolve", headers=HEADERS,
                         json={"playlist_url": "https://www.youtube.com/playlist?list=PLAYLIST_ID",
                               "limit": 100}).json()

ids = [v["video_id"] for v in playlist["items"]]
batch = requests.post(f"{API}/batch", headers=HEADERS, json={
    "video_ids": ids,
    "format": {"timestamp": True, "paragraphs": True},
}).json()

# Large batches run async: poll GET /batch/{batch_id} until status is "completed"
for r in batch.get("results", []):
    if r["status"] == "completed":
        for p in r["data"]["transcript"].get("paragraphs") or []:
            store(video_id=r["video_id"], start=p["start"], text=p["text"])  # your vector DB

How it makes money

Charge per report for brands that vet occasionally, and sell seat-based plans to agencies and creator marketplaces that vet every week. The price is easy to justify against the cost of one failed sponsorship.

Mistakes to avoid

• ✕Flagging without evidence: every risk needs a quote and timestamp so a human can judge context.
• ✕Treating all profanity as equally risky: let each brand set its own thresholds.
• ✕Analyzing only recent videos: competitor deals often appear further back.

Questions about this idea