JimmySadek/video-fetcher-to-markdown

#!/usr/bin/env bash

"${TMPDIR:-/tmp}"/*|/tmp/*|/var/folders/*/T/*) rm -rf -- "$probe_root" ;;

Read the [v2.0.0 release notes](https://github.com/JimmySadek/video-fetcher-to-markdown/releases/tag/v2.0.0)

title: "Obsidian: The King of Learning Tools (FULL GUIDE + SETUP)" channel: "Odysseas" url: "https://www.youtube.com/watch?v=hSTy_BInQs8"

channel: "Odysseas" url: "https://www.youtube.com/watch?v=hSTy_BInQs8" video_id: "hSTy_BInQs8"

url: "https://www.youtube.com/watch?v=hSTy_BInQs8"

| Field    | Value | |----------|-------| | URL      | https://www.youtube.com/watch?v=hSTy_BInQs8 |

|----------|-------| | URL      | https://www.youtube.com/watch?v=hSTy_BInQs8 | | Channel  | Odysseas |

| URL      | https://www.youtube.com/watch?v=hSTy_BInQs8 |

The YAML frontmatter makes a collection queryable through tools such as [Dataview](https://github.com/blacksmithgu/obsidian-dataview), while the Markdown

[Dataview](https://github.com/blacksmithgu/obsidian-dataview), while the Markdown

git clone https://github.com/JimmySadek/video-fetcher-to-markdown.git

python3 scripts/fetch_transcript.py "https://youtu.be/VIDEO_ID"

python3 scripts/fetch_media.py --check-deps python3 scripts/fetch_media.py --hint "Claude, HyperFrames" -- "https://www.tiktok.com/@user/video/123"

python3 scripts/fetch_media.py --hint "Claude, HyperFrames" -- "https://www.tiktok.com/@user/video/123"

[Credential-related source omitted]

[Kit to Clip](https://github.com/JimmySadek/kit-to-clip) turns a brand kit into finished, on-brand videos. Say

[Credential-related source omitted]

The [upstream caption library](https://github.com/jdepoix/youtube-transcript-api)

and link. Keep the skill name `youtube-fetcher` for the same reason: `npx skills update` re-adds each install by that name. See [GitHub releases](https://github.com/JimmySadek/video-fetcher-to-markdown/releases)

re-adds each install by that name. See [GitHub releases](https://github.com/JimmySadek/video-fetcher-to-markdown/releases) for tagged versions and upgrade notes.

[GitHub releases](https://github.com/JimmySadek/video-fetcher-to-markdown/releases)

# SKILL_DIR is the directory containing this SKILL.md python3 "$SKILL_DIR/scripts/fetch_transcript.py" -- "https://youtu.be/VIDEO_ID"

python3 "$SKILL_DIR/scripts/fetch_transcript.py" -- "https://youtu.be/VIDEO_ID"

# Note with transcript and (for videos up to 3 minutes) a contact sheet of frames python3 "$SKILL_DIR/scripts/fetch_media.py" -- "https://www.tiktok.com/@user/video/123"

python3 "$SKILL_DIR/scripts/fetch_media.py" -- "https://www.tiktok.com/@user/video/123"

2. **Browser fallback, no login.** If you have a browser tool, open the page, run the    bundled `scripts/browser_media_links.js` in it (it returns the best `audio` and    `video` links, title and description), download both at once with `c

bundled `scripts/browser_media_links.js` in it (it returns the best `audio` and    `video` links, title and description), download both at once with `curl -L -o`    (the links are signed and expire within hours), then:

`video` links, title and description), download both at once with `curl -L -o`

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

try { const e = new URL(u).searchParams.get('efg'); return JSON.parse(atob(e.replace(/-/g, '+').replace(/_/g, '/'))); }

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

#!/usr/bin/env python3

[Credential-related source omitted]

url = f"https://www.youtube.com/watch?v={video_id}"

url = f"https://www.youtube.com/oembed?url=https://www.youtube.com/watch?v={video_id}&format=json"     try:         resp = requests.get(url, timeout=timeout)

try:         resp = requests.get(url, timeout=timeout)         resp.raise_for_status()

resp = requests.get(url, timeout=timeout)

def timestamp_link(seconds: float, video_id: str) -> str:     return f"[{format_timestamp(seconds)}](https://www.youtube.com/watch?v={video_id}&t={max(0, int(seconds))}s)"

return f"[{format_timestamp(seconds)}](https://www.youtube.com/watch?v={video_id}&t={max(0, int(seconds))}s)"

video_url = f"https://www.youtube.com/watch?v={video_id}"

file=sys.stderr,                 )             transcript = selected.fetch()

)             transcript = selected.fetch()             if not transcript or not any(

transcript = selected.fetch()

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

curl -s "https://skills.sh/api/v1/skills/audit/jimmysadek/youtube-fetcher-to-markdown/youtube-fetcher"

[Credential-related source omitted]

[Credential-related source omitted]

[v1.2.0](https://github.com/JimmySadek/youtube-fetcher-to-markdown/releases/tag/v1.2.0)

was published on 2026-09-05 at `7401ded6c6195f38db962c95a3cc0a92a96bf42a`, through [PR #4](https://github.com/JimmySadek/youtube-fetcher-to-markdown/pull/4). All Linux, Windows, and macOS jobs passed in

through [PR #4](https://github.com/JimmySadek/youtube-fetcher-to-markdown/pull/4).

All Linux, Windows, and macOS jobs passed in [PR CI](https://github.com/JimmySadek/youtube-fetcher-to-markdown/actions/runs/33985647882) and [post-merge CI](https://github.com/JimmySadek/youtube-fetcher-to-markdown/actions/runs/33985725827)

[PR CI](https://github.com/JimmySadek/youtube-fetcher-to-markdown/actions/runs/33985647882)

and [post-merge CI](https://github.com/JimmySadek/youtube-fetcher-to-markdown/actions/runs/33985725827).

The [legacy mirror check](https://github.com/JimmySadek/youtube-fetcher-to-markdown/actions/runs/33985725846)

</style> <script type="module">   import mermaid from "https://cdn.jsdelivr.net/npm/mermaid@11.16.0/dist/mermaid.esm.min.mjs";

<script type="module">   import mermaid from "https://cdn.jsdelivr.net/npm/mermaid@11.16.0/dist/mermaid.esm.min.mjs";   const css = getComputedStyle(document.documentElement);

import mermaid from "https://cdn.jsdelivr.net/npm/mermaid@11.16.0/dist/mermaid.esm.min.mjs";

<header>     <h1>Plan: YouTube Fetcher v1.1 Release and Distribution</h1>     <p><strong>Historical v1.1 record.</strong> Engineering continues in the <a href="../README.md">current README</a> and <a href="youtube-fetcher-reliability-and-la

<h1>Plan: YouTube Fetcher v1.1 Release and Distribution</h1>     <p><strong>Historical v1.1 record.</strong> Engineering continues in the <a href="../README.md">current README</a> and <a href="youtube-fetcher-reliability-and-languages.md">v

<p><strong>Historical v1.1 record.</strong> Engineering continues in the <a href="../README.md">current README</a> and <a href="youtube-fetcher-reliability-and-languages.md">v1.2 reliability and language record</a>. Unless explicitly update

<dt>Agent</dt><dd>Codex</dd>         <dt>Session</dt><dd>Current Codex desktop planning task</dd>         <dt>Back references</dt><dd><a href="../README.md">README</a>, <a href="../SKILL.md">skill contract</a>, <a href="../scripts/fetch_tra

<dt>Session</dt><dd>Current Codex desktop planning task</dd>         <dt>Back references</dt><dd><a href="../README.md">README</a>, <a href="../SKILL.md">skill contract</a>, <a href="../scripts/fetch_transcript.py">runtime script</a>, <a hr

<dt>Back references</dt><dd><a href="../README.md">README</a>, <a href="../SKILL.md">skill contract</a>, <a href="../scripts/fetch_transcript.py">runtime script</a>, <a href="https://skills.sh/jimmysadek/youtube-fetcher-to-markdown/youtube-

<dt>Forward references</dt><dd>The release, CodeGuilds issue URLs, and <a href="https://github.com/VoltAgent/awesome-agent-skills/pull/869">VoltAgent pull request #869</a> are recorded below. Future directory submissions still require their

<ul>         <li><code class="status">[x]</code> The full deterministic regression suite passed on all tested Python versions.</li>         <li><code class="status">[x]</code> <code>python3 scripts/fetch_transcript.py 'https://youtu.be/hSTy

<li><code class="status">[x]</code> The full deterministic regression suite passed on all tested Python versions.</li>         <li><code class="status">[x]</code> <code>python3 scripts/fetch_transcript.py 'https://youtu.be/hSTy_BInQs8' --st

<li><code class="status">[x]</code> <code>python3 scripts/fetch_transcript.py 'https://youtu.be/hSTy_BInQs8' --stdout --no-description</code> passed end to end without saving a transcript.</li>

<li><code class="status">[x]</code> Renamed the local baseline branch to <code>main</code>, tracked <code>origin/main</code>, and refreshed <code>origin/HEAD</code>. The separate Claude worktree remains at <code>d07e035</code> with its untr

<li><code class="status">[x]</code> Raw <code>main</code> and compatibility <code>master</code> README and banner URLs return HTTP 200; the live CodeGuilds banner loaded at 1536 × 1024.</li>         <li><code class="status">[x]</code> Pull 

<li><code class="status">[x]</code> Pull request <a href="https://github.com/JimmySadek/youtube-fetcher-to-markdown/pull/2">#2</a> merged cleanly at <code>57ae480</code> after Python 3.8 and 3.13 checks passed.</li>

<li><code class="status">[x]</code> Post-merge CI run <a href="https://github.com/JimmySadek/youtube-fetcher-to-markdown/actions/runs/30777209683">30777209683</a> and legacy-sync run <a href="https://github.com/JimmySadek/youtube-fetcher-to

<li><code class="status">[x]</code> Annotated tag <code>v1.1.0</code> and the <a href="https://github.com/JimmySadek/youtube-fetcher-to-markdown/releases/tag/v1.1.0">GitHub Release</a> point to merged commit <code>57ae480</code>.</li>

<p>CodeGuilds work needs its own external-action approval and may proceed in parallel with directory research after v1.1 is public.</p>       <ul>         <li><code class="status">[x]</code> Reported the published npm CLI package-path failu

<ul>         <li><code class="status">[x]</code> Reported the published npm CLI package-path failure as <a href="https://github.com/xdevsapps/codeguilds-cli/issues/1">codeguilds-cli #1</a> with the isolated stack trace and verified one-line

<li><code class="status">[x]</code> Reported the published npm CLI package-path failure as <a href="https://github.com/xdevsapps/codeguilds-cli/issues/1">codeguilds-cli #1</a> with the isolated stack trace and verified one-line source corre

<li><code class="status">[x]</code> Corrected the clipped tagline, refreshed the claimed listing to the released v1.1 README, published its v1.1.0 version record, and set the truthful <code>npx skills</code> install command.</li>         <l

<li><code class="status">[x]</code> Re-tested <code>codeguilds info</code> and <code>codeguilds install --project</code> inside disposable home and project directories. The registry reports v1.1.0, but installation falsely succeeds after co

<li><code class="status">[x]</code> Reported the incomplete multi-file skill behavior and ignored project scope as <a href="https://github.com/xdevsapps/codeguilds-cli/issues/2">codeguilds-cli #2</a>; current upstream types can describe onl

<li><code class="status">[x]</code> <strong>VoltAgent preflight:</strong> captured 191 stars, 15 forks, and 118 skills.sh installs; re-read the live policy; found no duplicate YouTube Fetcher entry; verified the v1.1 release and canonical r

<li><code class="status">[x]</code> <strong>Existing discovery:</strong> skills.sh remains live at 118 installs. CodeGuilds shows zero web downloads and one CLI record from the 2026-08-03 isolated verification; that record is not evidence o

<li><code class="status">[x]</code> <strong>Wave 1 — submitted:</strong> opened <a href="https://github.com/VoltAgent/awesome-agent-skills/pull/869">VoltAgent pull request #869</a> under Productivity and Collaboration. The link-only entry u

<li><code class="status">[]</code> <strong>Conditional human channel:</strong> the live <a href="https://github.com/travisvn/awesome-claude-skills/blob/main/CONTRIBUTING.md">Travis policy</a> prohibits AI-generated or AI-submitted proposals

<li><code class="status">[x]</code> <strong>Awesome Claude Code decision — hold:</strong> the <a href="https://github.com/hesreallyhim/awesome-claude-code/issues/new?template=recommend-resource.yml">live human form</a> was inspected. It req

<li><code class="status">[]</code> <strong>Optional Wave 2 — staged:</strong> <a href="https://www.awesomeskills.dev/en/submit">AwesomeSkills.dev</a> accepts a GitHub URL and publishes instantly. Its form was inspected, but submission remai

<tr><td>Move from <code>master</code> to <code>main</code></td><td>Delivered safely</td><td><code>main</code> is canonical; the automatically fast-forwarded <code>master</code> compatibility branch preserves raw legacy links without divergi

<tr><td>CodeGuilds verification</td><td>Partially delivered</td><td>Listing, banner, tagline, README, version, and correct Skills CLI command are live. CodeGuilds' own multi-file installer remains upstream-blocked in issues #1 and #2. Its o

<tr><td>Selective library distribution</td><td>Wave 1 submitted</td><td><a href="https://github.com/VoltAgent/awesome-agent-skills/pull/869">VoltAgent pull request #869</a> is open. Travis remains human-authored only. Awesome Claude Code is

<ul>       <li>VoltAgent's prior Strategic Partner rejection was specifically about missing real community usage; YouTube Fetcher's present evidence addresses that concern but does not guarantee acceptance.</li>       <li>VoltAgent showed 6

<li>VoltAgent's prior Strategic Partner rejection was specifically about missing real community usage; YouTube Fetcher's present evidence addresses that concern but does not guarantee acceptance.</li>       <li>VoltAgent showed 65 open pull

<li>VoltAgent showed 65 open pull requests immediately after YouTube Fetcher became <a href="https://github.com/VoltAgent/awesome-agent-skills/pull/869">pull request #869</a>. Its link-only, cross-agent policy and mature-usage requirement m

<p><strong>2026-08-03 — local release candidate:</strong> Phases 1–4 completed in feature commit <code>f3a5028</code>. Deterministic tests, Python 3.8/3.13 compatibility, skill validation, formatting and shell checks, isolated installation,

<p><strong>2026-08-03 — v1.1 publication:</strong> Jimmy approved the remote gate and later authorized the necessary CodeGuilds correction. GitHub's default moved to <code>main</code>; a synchronized <code>master</code> compatibility branch

<p><strong>2026-08-03 — CodeGuilds truth:</strong> The claimed page now shows the released tagline, README, correct Skills CLI command, banner, and version 1.1.0. An isolated CodeGuilds install proved its current single-file skill model can

<p><strong>2026-08-03 — end-to-end distribution plan:</strong> Live policy, queue, duplicate, provenance, and usage research confirms that no additional library submission has occurred. VoltAgent is the recommended first link-based PR; Trav

<p><strong>2026-08-04 — browser-led distribution execution:</strong> Jimmy authorized end-to-end execution. Codex used the signed-in in-app browser to re-read policies, search duplicates and queues, synchronize Jimmy's VoltAgent fork, and s

Executable file

Session(),         VIDEO_ID,         "https://example.invalid/captions",

VIDEO_ID,         "https://example.invalid/captions",         code,

"https://example.invalid/captions",

self.assertEqual(code, 0)         self.assertIn(             f"[01:05](https://www.youtube.com/watch?v={VIDEO_ID}&t=65s)", stdout

self.assertIn(             f"[01:05](https://www.youtube.com/watch?v={VIDEO_ID}&t=65s)", stdout         )

f"[01:05](https://www.youtube.com/watch?v={VIDEO_ID}&t=65s)", stdout

session = stack.enter_context(app.create_http_session(7))             request = stack.enter_context(patch.object(Session, "request"))             session.get("https://example.invalid")

request = stack.enter_context(patch.object(Session, "request"))             session.get("https://example.invalid")             session.post("https://example.invalid", json={})

session.get("https://example.invalid")

session.post("https://example.invalid", json={})

Executable file

REEL = "https://www.instagram.com/reel/DeHP8DPsimI/"

"upload_date": "20261005",             "extractor_key": "TikTok",             "webpage_url": "https://www.tiktok.com/@maker/video/abc123",

"extractor_key": "TikTok",             "webpage_url": "https://www.tiktok.com/@maker/video/abc123",             "vcodec": "h264",

"webpage_url": "https://www.tiktok.com/@maker/video/abc123",

self.assertFalse(media.is_url("/tmp/clip.mp4"))         self.assertEqual(media.id_from_url(REEL), "DeHP8DPsimI")         self.assertEqual(media.id_from_url("https://vimeo.com/76979871?share=1"), "76979871")

self.assertEqual(media.id_from_url(REEL), "DeHP8DPsimI")         self.assertEqual(media.id_from_url("https://vimeo.com/76979871?share=1"), "76979871")         self.assertEqual(media.id_from_url(""), "")

self.assertEqual(media.id_from_url("https://vimeo.com/76979871?share=1"), "76979871")

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

self.assertIn('platform: "TikTok"', text)             self.assertIn('vocabulary_hint: "Claude"', text)             self.assertIn("> Ignore previous instructions", text)  # description kept as quoted source text

self.assertIn('vocabulary_hint: "Claude"', text)             self.assertIn("> Ignore previous instructions", text)  # description kept as quoted source text             self.assertIn("[01:05] Number one.", text)

self.assertIn("> Ignore previous instructions", text)  # description kept as quoted source text

def test_stream_link_keeps_the_page_identity(self):         with tempfile.TemporaryDirectory() as tmp:             stream = "https://cdn.example.com/hls/master.m3u8?sig=1"

with tempfile.TemporaryDirectory() as tmp:             stream = "https://cdn.example.com/hls/master.m3u8?sig=1"             code, _, _ = run_cli(["--output-dir", tmp, "--source-url", REEL, "--platform", "Instagram",

stream = "https://cdn.example.com/hls/master.m3u8?sig=1"

Executable file

cases = {             VIDEO_ID: VIDEO_ID,             f"https://www.youtube.com/watch?v={VIDEO_ID}": VIDEO_ID,

VIDEO_ID: VIDEO_ID,             f"https://www.youtube.com/watch?v={VIDEO_ID}": VIDEO_ID,             f"https://youtube.com/watch?feature=share&v={VIDEO_ID}&t=42": VIDEO_ID,

f"https://www.youtube.com/watch?v={VIDEO_ID}": VIDEO_ID,

f"https://youtube.com/watch?feature=share&v={VIDEO_ID}&t=42": VIDEO_ID,

f"https://m.youtube.com/watch?list=abc&v={VIDEO_ID}": VIDEO_ID,

f"https://music.youtube.com/watch?v={VIDEO_ID}": VIDEO_ID,

f"https://youtu.be/{VIDEO_ID}?si=abc": VIDEO_ID,

f"https://youtube.com/embed/{VIDEO_ID}": VIDEO_ID,

f"https://www.youtube-nocookie.com/embed/{VIDEO_ID}": VIDEO_ID,

f"https://youtube.com/shorts/{VIDEO_ID}?feature=share": VIDEO_ID,

f"https://youtube.com/live/{VIDEO_ID}": VIDEO_ID,

def test_rejects_lookalike_hosts_and_invalid_ids(self):         cases = [             f"https://youtube.com.example.org/watch?v={VIDEO_ID}",

cases = [             f"https://youtube.com.example.org/watch?v={VIDEO_ID}",             f"https://notyoutube.com/watch?v={VIDEO_ID}",

f"https://youtube.com.example.org/watch?v={VIDEO_ID}",

f"https://notyoutube.com/watch?v={VIDEO_ID}",

"https://youtube.com/watch?feature=share",

"https://youtu.be/too-short",

Executable file