YouTube Analytics Fetcher
Data & Research webFetches the channel owner's private YouTube analytics (retention, traffic source, CTR/impressions, subscriber source, watch time) and returns a per-video normalized JSON data table.
Live output preview
A plan is required to view this content
Choose a plan to access input format, sample outputs, and live previews.
View Plans →About the skill
YouTube Analytics Fetcher
Fetches the channel owner's private YouTube Analytics data (not public) directly via the YouTube Analytics API v2. While youtube-data-fetcher only provides public metrics, this skill accesses owner-only growth and retention signals: audience retention curve, traffic source distribution (YT_SEARCH / SUGGESTED_VIDEO / BROWSE / EXTERNAL), impressions and CTR, subscriber gain/loss source, watch time, and per-video performance.
When is it used: When deciding on a content revision (hook, thumbnail, title, SEO) by measuring which video truly drives subscribers, where viewers drop off after the hook, where traffic comes from, and the thumbnail/title CTR. The retention + CTR + traffic trio are the most valuable signals for content optimization.
Output: Per the mode selection (summary / by_video / traffic / retention / reach / subscribers), a normalized row table — as a JSON dataset. The sample data.json contains a per-video performance breakdown of roughly the last 12 months: views, watch time, average view percentage, CTR, impressions, net subscriber change, and the dominant traffic source.
Important: Requires an OAuth 2.0 refresh_token (does not work with a service account). Read-only — it never publishes or modifies anything. On new/low-viewership channels some metrics may come back empty; in that case "insufficient data" is reported and nothing is fabricated.
How do I use this skill?
Upload the youtube-analytics-fetcher.zip you downloaded as-is — no packaging needed, the format is already correct (folder at root).
- Open Settings → Customize → Skills
- Upload → select the
youtube-analytics-fetcher.zipyou downloaded - Claude reads
SKILL.md; the name + description appear. Ready ✅
Scripts run in Anthropic's code-execution environment (sandbox) — not on your machine.