| name | youtube-to-links |
| description | Import a YouTube video or full playlist into apps/media/content/links/ as Keystatic MDX entries. Use when the user gives a YouTube URL and asks to add it (or all videos in it) to the media app's links collection. Channel-aware: videos from the Solana Foundation channel (@SolanaFndn) get tagged automatically. |
YouTube → apps/media/content/links
Turn a YouTube URL into one MDX file per video under
apps/media/content/links/, matching the schema in
apps/media/keystatic.config.tsx and the formatting of existing entries like
apps/media/content/links/catherine-gu-sdp-interview-fireblocks.mdx.
When to use
The user provides a YouTube URL (single video, short, or playlist) and asks to
import / add / save it to the links collection.
youtube.com/watch?v=…, youtu.be/…, youtube.com/shorts/… → one MDXyoutube.com/playlist?list=… → one MDX per video in the playlist
How to use
Run the helper script. It reads YOUTUBE_DATA_API_KEY (or YOUTUBE_API_KEY)
from the environment, falling back to apps/media/.env.local:
python3 skills/youtube-to-links/import_youtube.py "<youtube-url>"
Multiple URLs are accepted and de-duped. Existing files (matched by slug) are
left untouched, so re-runs are safe.
What gets written
Frontmatter mirrors the existing links schema:
---
title: "<video title>"
url: "https://www.youtube.com/watch?v=<id>"
linkType: video
description: >
<video title>
source: "YouTube"
publishedAt: <video publishedAt, ISO 8601 with ms>
tags:
- tag: solana-foundation
featured: false
---
<video title>
Fixed values (per the project owner):
source: "YouTube"linkType: videofeatured: falsedescription and MDX body are the video title (matches the Catherine Gu
example). No category is added by default — fill in via Keystatic if needed.
Channel → tag mapping
| Channel | Tag applied |
|---|
Solana (@SolanaFndn, channel id UC9AdQPUe4BdVJ8M9X7wxHUA) | solana-foundation |
Add new mappings in import_youtube.py (is_solana_foundation /
SOLANA_CHANNEL_* constants) if more channels need tagging later.
Slug
The slug matches Keystatic's behavior (lowercase, non-alphanumerics → -,
trimmed and truncated to 80 chars) so the filename matches what Keystatic would
have generated had the entry been authored in the CMS.
Duplicate check
After writing, the script scans every .mdx in apps/media/content/links/,
extracts the url: field, and groups by YouTube video id (falling back to the
raw URL for non-YouTube links). Any group with more than one file is printed as
a duplicate group, e.g.:
Found 1 duplicate group(s) in apps/media/content/links:
yt:UW8qaI5SbEY
- apps/media/content/links/accelerate-usa-2026-anatoly-yakovenko.mdx
- apps/media/content/links/some-other-file-pointing-at-the-same-video.mdx
If you see this, decide which file to keep and delete the other — the script
does not auto-delete.
After running
- The script prints
written / skipped for each video, a final summary, and
the duplicate-check result.
- Spot-check one of the generated files against
apps/media/content/links/catherine-gu-sdp-interview-fireblocks.mdx.
- If you want categories or extra tags, add them in Keystatic — the script
leaves both empty by default (except the channel-based tag above).
