Post

DevTools to .ts: My Full Walkthrough of Building M3U8-Probe

build a dual-script video extractor from raw .m3u8 or direct .ts links, with async download, DevTools tactics, and ts-to-mp4 merging. Plus: guide for using link_dl.py & m3u8_dl.py.

Posted Updated
By phruit
12 min read
DevTools to .ts: My Full Walkthrough of Building M3U8-Probe

Well, we’re all big fans of free comedies and shows… so why not grab a bucket of popcorn and let Python do the dirty work for us?

This is a detailed guide for M3U8-Probe. Hope this helps! And have fun! 🐈

Heads-up: These are script-based tools for now - but I’m planning to wrap them into CLI tools in the future, once they prove rock solid.

🧪 Tested Sites (So far)

These scripts have successfully worked on the following real-world sites(all for phree shows!). Feel free to test more.

Know more sites that work? PRs or issues welcome!

https://streamwatch.online/media/tmdb-movie-574475-Final%20Destination%20Bloodlines
https://ww4.fmovies.co/film/phineas-and-ferb-season-5-1630859174/
https://moviexfilm.com/on-swift-horses-2/
https://www.bttwo.me/
https://xiaoyakankan.com/post/f9905e1327.html?vod=190_11408-9

P.S. Rumor has it… these scripts also work on some spicy sites. But don’t ask me how I know.

🧪 Which Script Should You Use?

💻 Step 1: Can you find .m3u8 files in DevTools?

(e.g. https://play.com/20240701/1rIWCjMw/2000kb/hls/index.m3u8)

🌐 Step 2: Can you find sequential segment links in DevTools?

Look at how the numbers increase in the URLs:

Segment StyleExamplePatternScript to Use
Increasing at the endhttps://play.gotomymv.life/.../000.ts → .../145.ts.../000.ts → .../001.ts → …link_dl.py
Increasing in the middlehttps://.../seg-1-v1-a1.ts?... → .../seg-177-v1-a1.ts?...seg-1-... → seg-2-... → …link_dl.py

If you spot this kind of numbering - anywhere in the link - and it increases chunk by chunk, that’s your green light for link_dl.py.

  • Yes → Use link_dl.py

  • No → Sorry bruh, these scripts won’t help here. Might need to write a custom one or inspect more.

🎬 m3u8_dl.py - Extracting .m3u8 Files

🚀 What m3u8_dl.py Does

Async segment scanner + downloader + merger

m3u8_dl.py is for .m3u8 playlist-based video downloads.
It works best when the playlist contains a clean list of video chunks.

📖 Step 1: Read lines

Reads your .m3u8 playlist line by line, sniffs out only the good stuff (a.k.a. segment filenames).

Needs a full URL to the .m3u8 file, so it can reconstruct (segments’) full URLs like a smooth operator.

🍿 Step 2: Download & Merge

Each segment is downloaded asynchronously using aiohttp. Then it’s auto-merged into a single .ts file. Just play and chill.

  1. Open DevTools Hit F12 while the video is loading. If no network activity shows, hit Ctrl + R to reload.

  2. Filter Requests Go to the Network tab and filter by XHR and Media (you can also add HTML and JS in case some sites are tricky).

  3. Find the .m3u8 file Look for a file ending in .m3u8 - there could be various names, like index.m3u8, playlist.m3u8, or even something sneaky without “m3u8” in the name at all. Click it, and check the Response tab.
    You wanna see a full m3u8 playlist like this:

    #EXTM3U
#EXT-X-VERSION:3
#EXT-X-TARGETDURATION:4
#EXT-X-PLAYLIST-TYPE:VOD
#EXT-X-MEDIA-SEQUENCE:0
#EXTINF:2.085,
/20240701/1rIWCjMw/2000kb/hls/Bhr5jO2P.jpg
#EXTINF:3.17,
/20240701/1rIWCjMw/2000kb/hls/k5gKM4Iw.jpg
...
#EXT-X-ENDLIST

    Though these segments end with .jpg, they’re actually .ts video chunks in disguise.
    You can expose their true identity using your good ol’ friend curl.

    Pick one, manually append it to the base URL, and run:

    1

    curl -I 'https://play.modujx12.com/20240701/1rIWCjMw/2000kb/hls/Bhr5jO2P.jpg'

    If the response looks like this (instead of returning image metadata), then you’ve caught a .ts file wearing a .jpg mask:

    Accept-Ranges: bytes
Access-Control-Allow-Headers: X-Requested-With
Access-Control-Allow-Methods: POST, GET, OPTIONS
Access-Control-Allow-Origin: *
Content-Disposition: attachment; filename="Bhr5jO2P.ts"
Content-Length: 324676
Content-Type: application/octet-stream
Date: Wed, 18 Jun 2025 06:34:19 GMT
Etag: "66831471-4f444"
Last-Modified: Mon, 01 Jul 2024 20:41:21 GMT
...

    Some CDN streaming services do this to obfuscate or bypass firewalls/caching rules.

  4. Right click that request, Copy Value -> Copy URL, that is your remote_m3u8_url.

📂 Example .m3u8 file

Here’s an example .m3u8 file.

Toward the end, you’ll see lines like:

#EXT-X-DISCONTINUITY
#EXT-X-KEY:METHOD=NONE
#EXTINF:3.333,
/20250612/AOraqYHt/8984kb/hls/2LZ5mvIk.jpg
...
#EXT-X-ENDLIST

Now compare that with more typical chunks like:

#EXTINF:2.085,
/20240701/1rIWCjMw/2000kb/hls/YTXeysd2.jpg
#EXTINF:0.542,
/20240701/1rIWCjMw/2000kb/hls/6msgPZUM.jpg

If you try downloading from the later chunk (/20250612/AOraqYHt/8984kb/hls/0gOMoF2r.jpg), you’ll often find… surprise! It’s an ad.

⚠️ Watch for #EXT-X-DISCONTINUITY

This tag means:

“Heads up! The next media chunk is from a different timeline or media source.”

If you see multiple #EXT-X-DISCONTINUITY tags in a playlist, they’re often marking:

  • The boundary between ads and show
  • Different chapters or transitions
  • Totally different video parts spliced together

That’s why get_segs() steps in-to filter out the trash and keep only the valid prefixes. Learn more in this part - Troubleshooting.

🧪 Run m3u8_dl.py - Download Using .m3u8 Playlist

Once you’ve got your remote_m3u8_url ready, you can use this script to download the full video.

📝 How to Use m3u8_dl.py

Edit the parameters directly under the if __name__ == "__main__": block in the script:

1
2
3
4
5
6
7
8
9
10
11

# Configuration section
    
remote_m3u8_url = "https://play.com/20240701/1rIWCjMw/2000kb/hls/index.m3u8"
output_file = "wrecked_s01e01.ts"

# Headers used for HTTP requests
headers = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36",
    "Referer": "https://example.com",
    "Origin": "https://example.com"
}

🧷 Parameter Breakdown for m3u8_dl.py

Set these values in the script under if __name__ == "__main__": before running.

  1. remote_m3u8_url - Remote full URL to the .m3u8 file (in DevTool).

    Example:

    1

     remote_m3u8_url = "https://play.com/20240701/1rIWCjMw/2000kb/hls/index.m3u8"

  2. output_file - Your Output Filename

    Make it cute or clean, e.g.:

    1

     output_file = "wrecked_s01e10.ts"

  3. headers - Request Headers for Downloading

    Usually, just tweak Referer and Origin based on what you saw in the browser.

    Examples:

    1
2
3
4
5

     headers = {
     "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36",
     "Referer": "https://www.bttwo.me",
     "Origin": "https://www.bttwo.me"
 }

    Peek at the actual segment request in DevTool → Headers tab → scroll to Request Headers to copy & adapt.

  4. (Optional) Retry Times in download_seg()

Sometimes the net gets cranky, or the server ghosts us mid-download.
So by default, each segment gets retried up to 3 times before giving up.

Just tweak this line in the function if you wanna change:

1

async def download_seg(session, sem, base_url, seg, idx, output_dir, max_retries=3):  # 👈 Change this "3" to any number you want

Run m3u8_dl.py

1
2
3
4
5
6
7
8

git clone https://github.com/kay-a11y/M3U8-Probe.git
cd M3U8-Probe

# 📦 Install dependencies
pip install -r requirements.txt

# 🚀 Run the script
python src/m3u8_dl.py

📋 Sample Output

2025-06-17 18:36:02,307 - INFO - base_url = https://play.modujx12.com/20240701/1rIWCjMw/2000kb/hls/
2025-06-17 18:36:08,707 - INFO - m3u8_file = /home/z3phyr/repos/GazeKit/hub-generalM3u8/data/wrecked_s01e01/playlist.m3u8
2025-06-17 18:36:08,710 - INFO - segs length = 623
💘 Downloading: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 623/623 [00:36<00:00,  7.15seg/s]

├── ...
├── README.md
├── wrecked_s01e10.ts   👈🏻 Your video is here!
├── data
│   └── wrecked_s01e10
│       └── index.m3u8  👈🏻 The playlist
└── src
    ├── link_dl.py
    └── m3u8_dl.py

🎬 link_dl.py - Extracting Request URLs

Let’s get your hands dirty with raw segment sniffing and real-time scraping.

🛠️ Find the Request URLs

  1. Open DevTools Hit F12 while the video is loading. If no network activity shows, hit Ctrl + R to reload.

  2. Filter Requests Go to the Network tab and filter by XHR and Media (you can also add HTML and JS in case some sites are tricky).

  3. Find the Request URLs Look for something like:

    https://play.gotomymv.life/ts/745bc3b1.../21633/000.png.ts

    or

    https://examplevideo.com/hls/videos/.../seg-1-v1-a1.ts?validfrom=...

    Check if the number in the URL increases between requests (e.g., 000.ts → 001.ts, seg-1.ts → seg-2.ts)

    You don’t need to find the last segment - as long as there’s continuity, the script will handle the rest.

🚀 What link_dl.py Does

link_dl.py is perfect for URL-based segment crawling, not .m3u8 playlist-based downloading.

📦 Step 1: Scanning

It auto-generates segment links by incrementing the number in the known request pattern.

🔄 Step 2: Wrapping into Tasks

Each URL is wrapped into tasks for async downloading.

🎞️ Step 3: Download & Merge

Downloaded segments are merged into a single final video.

🧪 Run link_dl.py - Download from Sequential Segment Links

Now that you’ve confirmed the request URL structure, it’s time to unleash link_dl.py and start downloading full videos like a stealthy ninja with a curl addiction.

📝 How to Use link_dl.py

  1. Edit the Config Below if __name__ == "__main__":

    Paste your target config like this:

    1
2
3
4
5
6
7

     base_url = "https://play.gotomymv.life/ts/40c26f81749767102/L3R5MzAvbWVpanUv5bCR5bm06LCi5bCU6aG_UzA1RTA4/20763/"
 output_file = "Sheldon_s05e08.ts"
 headers = {
     "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36",
     "Referer": "https://www.bttwo.me",
     "Origin": "https://www.bttwo.me"
 }

  2. Check the Segment Filename Pattern

    In the script (around line 65), you’ll see this key part:

    1

     seg_name = f"{i:03d}.png.ts"

    This tells the script to download segments like:

     000.png.ts, 001.png.ts, 002.png.ts ...

    Modify this line if your segment pattern is different, like:

    • f"seg-{i}-v1-a1.ts"
    • f"{i}.ts"
    • f"{i:03}.jpg"
    • …

    Match it exactly to what you saw in DevTools - See below.

🧷 Parameter Breakdown for link_dl.py

  1. seg_name

    At line 65, you’ll see:

    1

     seg_name = f"{i:03d}.png.ts"

    This is where the automation happens. Replace it to match the increasing part of your segment URL.

    Example:

    Full segment URL:

     https://examplevideo.com/hls/videos/202409/09/9999/720P_4000K_99.mp4/seg-1-v1-a1.ts?validfrom=1749397780&validto=1749404980&hash=qwerty

    The changing bit is seg-1, seg-2, seg-3…

    So your line should be:

    1

     seg_name = f"{i}-v1-a1.ts?validfrom=1749397780&validto=1749404980&hash=qwerty"

    Just keep the rest of the URL static, and replace the counter with {i} or {i:03d} depending on the pattern (zero-padded or not).

  2. base_url

    This is everything before the segment name starts changing.

    Using the same example:

     https://examplevideo.com/hls/videos/202409/09/9999/720P_4000K_99.mp4/seg-1-v1-a1.ts?validfrom=1749397780&validto=1749404980&hash=qwerty

    Here your base_url should be:

    1

     base_url = "https://examplevideo.com/hls/videos/202409/09/9999/720P_4000K_99.mp4/seg-"

  3. output_file

    Examples:

    1

     output_file = "Sheldon_s05e08.ts"

  4. headers - Request Headers for Downloading

    Usually, just tweak Referer and Origin based on what you saw in the browser.

    Examples:

    1
2
3
4
5

     headers = {
     "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36",
     "Referer": "https://example.com",
     "Origin": "https://example.com"
 }

    Peek at the actual segment request in DevTool → Headers tab → scroll to Request Headers to copy & adapt.

  5. (Optional) Retry Times in download_seg()

    Sometimes the net gets cranky, or the server ghosts us mid-download.
    So by default, each segment gets retried up to 3 times before giving up.

    Just tweak this line in the function if you wanna change:

    1

     async def download_seg(session, sem, seg_name, output_path, max_retries=3):  # 👈 Change this "3" to any number you want

1
2
3
4
5
6
7
8

git clone https://github.com/kay-a11y/M3U8-Probe.git
cd M3U8-Probe

# 📦 Install dependencies
pip install -r requirements.txt

# 🚀 Run the script
python src/link_dl.py

📋 Sample Output

1
2
3
4

🔍 Scanning segments: 148seg [01:39,  1.49seg/s]
💘 Downloading:  99%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████▍ | 146/148 [00:14<00:00, 16.33seg/s] 
2025-06-12 18:11:49,064 - WARNING - 💥 020.png.ts error:  (try 1/3)
💘 Downloading: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 148/148 [00:17<00:00,  8.41seg/s]

1
2
3
4
5
6
7
8

├── ...
├── README.md
├── Sheldon_s05e08.ts   👈🏻 Your video is here!
├── data
│   └── Sheldon_s05e08  👈🏻 Will be empty after merging, feel free to delete this
└── src
    ├── link_dl.py
    └── m3u8_dl.py

🎬 (Optional) Convert .ts to .mp4 with ffmpeg

When you’re done collecting all the juicy .ts segments into one big file, you might want to dress it up a bit - .mp4 style. Here’s how to glam it up with zero re-encoding:

1
2
3
4
5

# 🧰 Install ffmpeg (if not already)
sudo apt install ffmpeg

# 💃 Convert .ts to .mp4 without re-encoding
ffmpeg -i whole.ts -c copy final.mp4

This command simply re-wraps the .ts into .mp4 - no quality loss, no long wait. Just pure elegance.

🐞 Troubleshooting

Output video looks complete in size, but actually only plays a few seconds

Normally, we might use if not line.strip().startswith("#") to identify valid media segment lines in an .m3u8 file.
However, when ad chunks are present - typically following a #EXT-X-DISCONTINUITY tag - this method becomes unreliable.

If we stick to the original approach when processing .m3u8 files that include ad or junk segments, the resulting video might behave strangely.
For instance, you could end up with a clip that’s only a few seconds long but has a file size as large as a full video - like 10 seconds weighing in at 200 MB.

Here’s another method to extract valid media segment lines from an .m3u8 file (get_segs() in m3u8_dl.py):

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

legit_prefix = "/" + "/".join(base_url.split("/")[3:])  # e.g. /20240701/xxx/yyy/hls

with open(m3u8_file, 'r', encoding="utf-8") as f:
    for line in f:
        line = line.strip()
        if not line or line.startswith("#"):
            continue  # skip empty/comment lines

        # case 1: looks like /...jpg (actual .ts style)
        if line.startswith(legit_prefix):
            segs.append(line)

        # case 2: plain '000.ts'-style segment
        elif "/" not in line and line.endswith(".ts"):
            segs.append(line)

        # case 3: absolute URL with .ts extension
        elif line.startswith("http") and line.endswith(".ts"):
            segs.append(line)

This part first filters out empty lines and comments (those starting with #). Then, it categorizes segment lines into three types:

  1. Path-style segments that look like /...jpg but are actually .ts: These are validated using a legit_prefix, which is parsed from the remote .m3u8 URL (e.g. /20240701/xxx/yyy/hls). This helps exclude unrelated junk segments or ads.

  2. Plain numbered .ts segments: For example, 000.ts, 002.ts, etc. These are assumed to be valid by default.

  3. Absolute URLs ending in .ts: For example, https://astream.org/stream/613615ae/1080/index0.ts. These are also accepted as legit segments.

You can tweak the logic to adapt to different .m3u8 formats depending on the site’s layout or how the playlist is structured.

[Errno 111] Connection refused

2025-06-17 17:27:32,018 - ERROR - 💥 Failed to download M3U8: SOCKSHTTPSConnectionPool(host='play.modujx12.com', port=443): Max retries exceeded with url: /20240701/1rIWCjMw/2000kb/hls/index.m3u8 (Caused by NewConnectionError('<urllib3.contrib.socks.SOCKSHTTPSConnection object at 0x7172bf6fdcf0>: Failed to establish a new connection: [Errno 111] Connection refused'))

This error usually means your connection attempt was blocked or your proxy isn’t configured correctly.

If you’re using a proxy, double check your port settings.
Also run:

1

git config --global --list

to ensure your Git proxy settings aren’t interfering with Python requests.

See this section for troubleshooting tips: Solve Port Issues

Debug with logging into a file

Try to write logs right into a file, so you can keep receipts on every .ts, every 404, every little juicy fail or win.

To log into a file, just add the filename and filemode:

1
2
3
4
5
6
7
8
9

import logging
from datetime import date

logging.basicConfig(
    level=logging.DEBUG,
    format='%(asctime)s - %(levelname)s - %(message)s',
    filename=f"logs/{date.today()}.log",
    filemode='w'
)

💖 Support me with crypto or PayPal! 💘

💵 USDT (TRC20):
TJCANuMYSdgLKRKnpCtscXrS5NgDbBAvF9

🟠 Bitcoin (BTC):
bc1qrc9vhrrhnc9v9s9q9rjn24aj608j44p5hzsxft

Or support me on Ko-fi:

Support me on Ko-fi

Any amount helps me continue creating content 💬💻

🤖 tech, 🔒 Web Security
🎬 M3U8 🔎 DevTools 🐈‍⬛ GitHub 📺 video downloader ✍️ tutorial 🛠️ troubleshooting
This post is licensed under CC BY 4.0 by the author.