Connect your channel
This dashboard reads your channel as its owner over a read-only Google sign-in, so it can show the metrics that actually matter — watch time, retention curves, traffic sources, demographics and revenue — not just public view counts.
Scopes requested are read-only: youtube.readonly,
yt-analytics.readonly, yt-analytics-monetary.readonly. Nothing is ever written to your channel.
GOOGLE_CLIENT_ID and
GOOGLE_CLIENT_SECRET to your .env file, then restart the server.
See the README.md for the 5-minute Google Cloud setup.
AI summary — the story in one paragraph
Click Generate summary to have Claude read this window’s numbers and tell you the single most important takeaway and what to do next.
Trend over time
How people find you
Subscribers & loyalty
Engagement signals
Top videos — click one to read its retention curve
Audience retention
Who watched this one — age & gender, % of this video’s watch time
Compare with your own videos — same-list videos are the same type, so they’re the fair benchmark
Pick a list above to compare the videos in it against each other.
Who is watching
Age & gender (% of viewers) · real videos only (approx.)
Top geographies
Devices
Revenue
Subscriber growth — your audience over time, on any horizon
YouTube reports subscriber changes by whole days, so the total line is rebuilt backward from today’s live count — every point is anchored to the real number. A subscription is to the channel, so this count can’t exclude Shorts; the per-video attribution at the bottom covers real videos only.
Which videos win subscribers — real videos only, this horizon
Momentum — which videos are earning the views right now
Every real video that got views in the last 28 days, ranked by the window you pick — recent performance, not lifetime totals, so an old workhorse and a fresh upload compete on equal footing. Daily is the finest grain any YouTube API reports (hourly numbers exist only inside Studio), and analytics settle ~3 days back, so every window ends there. Shorts are excluded.
Views by vintage over time — new vs recent vs back catalogue, day by day
Each day’s views, split by how old the earning video was that day — released within the toggle’s window (“new”), in the x days before it (“recent”), or earlier (the back catalogue). A video migrates through the buckets as it ages, so this shows whether the channel runs on launches or on the library. The split re-cuts with the window toggle above.
Videos — every upload, with an AI read on what to repeat and fix
Pick a list to browse its videos. Each card shows the essentials — thumbnail, title, when it went out, and lifetime views — and can generate a one-paragraph Claude read of that video's notable numbers, what to replicate, and what to improve (measured against its same-list peers).
Choose a list above to load its videos.
Q&A prep — turn a video’s comments into the questions worth answering next
Pick a video to pull its comments. Then, when you’re ready, click Extract questions to have Claude cluster them into the questions worth answering in a follow-up — ordered by how many viewers raised each. Expand any question to see who asked and exactly how they put it.
Choose a list above to load its videos.
Questions to answer
Trends by type — how one format’s key stats move month over month
Each video sits in a playlist, and a playlist is a type (long-form explainers, a series). Pick one to see how that format’s monthly numbers have moved — combined across every real video in it. Shorts are excluded, so a Shorts-only playlist will read as empty.
Choose a type above to load its trends.
Demographics — how your audience is shifting over time
Each period is measured on its own, from April 2026. Gender and age are the share of signed-in viewers; nationality is the share of watch time by country. Real videos only — YouTube can’t split age & gender by content type, so those are approximated by view-weighting each real video’s audience (Shorts excluded). Read the shape of the trend, not any single point.
By video — which groups each video skews toward, vs. the rest of the list
Pick a list. For every video in it, this shows the age × gender groups it was relatively more popular with — i.e. that watched it more than they watched the list’s other videos on average. A quick way to see which topics pull which audiences. All videos covers the real uploads only; Shorts is its own entry, so the two kinds are never mixed.
Choose a list above to break its videos down by audience.
Retention curves — the last videos of one type, overlaid
Each line is one video’s audience-retention curve — the share still watching at each point through its length. Overlaying the most-recent real videos of one type shows whether your hook and pacing are improving, and which videos hold attention best. Shorts are excluded. Hover a video on the right to trace its curve; click to keep it selected.
Choose a list above to overlay its recent videos’ retention.