How to build a searchable archive of public meetings from YouTube

The problem it solves

Local journalists and residents want to know what officials said about a zoning change, a budget item or a school policy. Finding it means scrubbing through hours of video across many meetings. Agendas and minutes rarely capture the full discussion, and many local newsrooms no longer have reporters attending every meeting.

What to include in your first version

• ■Automatic ingestion of every meeting from the body's channel
• ■Speaker labels so you can search by official
• ■Topic and keyword search across years of meetings
• ■Alerts when a topic or address is discussed
• ■Shareable links to the exact moment in the recording

How to build it, step by step

1. 01

   Collect meeting videos

   Resolve the council's channel or meeting playlist and keep checking for new uploads.

2. 02

   Transcribe with speakers

   Use ASR with speakerLabels enabled, and speakerId with known names, so each passage is attributed to a speaker. Meetings are long, so use the estimate endpoint to preview cost.

3. 03

   Index by meeting and speaker

   Store paragraphs with meeting date, speaker and timestamp, and add full-text and semantic search.

4. 04

   Add alerts

   Let users save searches such as a street name or policy topic and email them when it comes up in a new meeting.

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

Why it's worth building

Local news outlets, advocacy groups and real estate or policy firms will pay for alerts and research tools. Civic projects often run on grants or offer a free public archive with paid alerts.

Mistakes to avoid

• ✕Relying on automatic captions for long meetings with many speakers: speaker labels make the archive far more useful.
• ✕Underestimating length: meetings run for hours, so estimate ASR credits before transcribing a backlog.
• ✕Publishing without context: link every quote to the recording so readers can hear it themselves.

Questions about this idea