Download videos from TLDV.io

tldv.py

from datetime import datetime
from os import system
import requests
import json

## Please install ffmpeg before running this script and make sure it's in your PATH
## brew install ffmpeg

## Please install requests before running this script
## pip3 install requests

## download video from tldv.io
##
## 1. Go to https://tldv.io/
## 2. Login
## 3. Go to the meeting you want to download
## 4. Copy the URL of the meeting
## 5. Open the developer tools (F12)
## 6. Go to the network tab
## 7. Refresh the page
## 8. Find the request to https://gw.tldv.io/v1/meetings/64145828ced74b0013d496ce/watch-page?noTranscript=true
## 9. Copy the auth token from the request headers
## 10. Run this script and paste the URL and auth token
## 11. python3 tldv.py

url = input("Please paste the URL of the meeting you want to download: ")

meeting_id = url.split("/")[-1]

print("\rFound meeting ID: ", meeting_id)

auth_token = input("Auth token: ")

data = requests.get(
    f"https://gw.tldv.io/v1/meetings/{meeting_id}/watch-page?noTranscript=true",
    headers={
        "Authorization": auth_token,
    },
)

try:
    response = json.loads(data.text)

    meeting = response.get("meeting", {})
    name = meeting.get("name", "No name")
    createdAt = meeting.get("createdAt", datetime.now())
    source = response.get("video", {}).get("source", None)

    date = datetime.strptime((createdAt), "%Y-%m-%dT%H:%M:%S.%fZ")
    normalised_date = date.strftime("%Y-%m-%d-%H-%M-%S")

    filename = f"{normalised_date}_{name}"
    filename_ext = ".mp4"

    command = f'ffmpeg -i {source} -c copy "{filename}.{filename_ext}"'

    json_filename = f'{filename}.json'

    with open(json_filename, "w") as f:
        f.write(data.text)

    print(command)

    print("Downloading video...")

    system(command)

except:
    print("Error encountered")
    print(data.text)

Works like a charm! I added a fork with parallel downloading in case you need to download and process multiple at a time. I obtained links with selenium, and then made a list of videos. The MAX_WORKERS are basically the number of parallel videos to download and process. This makes the process way much faster.

I've used this from javier-ortizt
I'm a newbie here, but I'm so glad I got it to work haha! Basically, I installed ffmpeg on my Mac using Homebrew (I followed this tutorial) and once it's set up, I ran the command

python3 tldv.py

Thanks so much!!

Note: If you feel ffmpeg is too slow when downloading .m3u8 videos chunk by chunk — try this faster version using N_m3u8DL-RE, which supports parallel segment downloading and auto-merging.

WIndows Version

---

Download N_m3u8DL-RE.exe:
👉 https://github.com/nilaoda/N_m3u8DL-RE/releases

---

Then run this script to download TLDV meetings more faster:

from datetime import datetime
from os import system
import requests
import json

url = input("Please paste the URL of the meeting you want to download: ")
meeting_id = url.split("/")[-1].split("?")[0]
print("\rFound meeting ID: ", meeting_id)

auth_token = input("Auth token: ")

data = requests.get(
    f"https://gw.tldv.io/v1/meetings/{meeting_id}/watch-page?noTranscript=true",
    headers={"Authorization": auth_token},
)

try:
    response = json.loads(data.text)
    meeting = response.get("meeting", {})
    name = meeting.get("name", "No name")
    createdAt = meeting.get("createdAt", datetime.now())
    source = response.get("video", {}).get("source", None)
    date = datetime.strptime((createdAt), "%Y-%m-%dT%H:%M:%S.%fZ")
    normalised_date = date.strftime("%Y-%m-%d-%H-%M-%S")
    filename = f"{normalised_date}_{name}"
    
    command = f'N_m3u8DL-RE.exe "{source}" --save-dir "." --save-name "{filename}" --thread-count 16 --del-after-done'
    print(command)
    print("Downloading video...")
    system(command)

    with open(f"{filename}.json", "w") as f:
        f.write(data.text)

except Exception as e:
    print("Error encountered")
    print("Exception:", str(e))
    print("Raw response:")
    print(data.text)

---

This script:

• Uses N_m3u8DL-RE to download in parallel (much faster than ffmpeg)

Created this dynamic script if anyone requires:

https://github.com/Cramraika/tldv_downloader/