Analytics
The “Analytics” section shows how your content performs — likes, comments, views and engagement across all connected platforms. And a separate “Agent analytics” page helps you keep track of how much and how well your AI agents are working. All numbers are calculated from real data of published posts, with no rough estimates.
Metrics overview
Open “Analytics” in the side menu. At the top there’s a period switcher: 7, 30 or 90 days. All the cards and charts on the page are recalculated for the selected window.
Five key metrics:
| Metric | What it shows |
|---|---|
| Likes | Total likes across all posts for the period |
| Comments | Total comments for the period |
| Views | Total views for the period |
| ER | Engagement rate — average engagement per post |
| Published | How many posts went out during the period |
Engagement trend
The “Engagement trend” chart shows views, likes and comments by day for the selected period. It makes it easy to see which days “took off” and to match peaks with specific posts. Hover over any point — a tooltip appears with the exact numbers for that day.
Breakdown by platform
The “By platform” block shows how many posts went to each connected social network — platforms are sorted by count. It’s a quick way to spot an imbalance: for example, if everything goes to Telegram while Instagram sits idle.
Best time to publish
The “Best time” heatmap is a grid of day of week × hour. The more saturated a cell, the higher the average engagement of posts published in that slot. The map is built from the real engagement of your published posts, not from generic “one-size-fits-all” recommendations.
Below the map there are the top 5 slots: the best combinations of day and hour, ranked by average engagement. The first slot is marked with a star.
AI insight
At the top of the page Repost21 shows a card with an AI insight — it takes your best slot from the heatmap and suggests a publishing window along with its engagement figure. The “Schedule” button takes you straight to creating a post for that time. The insight appears once you’ve accumulated data from at least a few published posts.
Export to CSV
The CSV button in the header exports the trend data for the selected period. The file opens correctly in Excel, including Cyrillic — handy for client reports or your own pivot tables.
Agent analytics
The “Agent analytics” page (in the “AI agents” section) is a separate dashboard for your agents’ work. It has the same 7/30/90 days period switcher and its own “Export” to CSV button.
Key metrics:
| Card | What it shows |
|---|---|
| Agents | How many agents are created and enabled |
| Runs | Number of runs for the period and the percentage of successful ones |
| Tokens | How many tokens were spent (input and output) |
| Spent | AI spend for the period and the daily average |
| Errors | Share of failed runs |
7-day trend
The chart shows agent runs by day over the past week: today is highlighted in a separate color, and below the chart there’s the total, the average and the spend, as well as the week-over-week delta (an up or down arrow) — so you can immediately see whether your agents sped up or slowed down.
Agent table
Each agent in the table comes with an avatar, role and the model (its “brain”) it uses. For each agent you can see:
- the number of runs and a mini activity bar relative to the busiest agent;
- tokens spent and the cost;
- when the agent last ran.
The table is sorted by runs — the most active agents are at the top. Clicking a row takes you to the agent’s card, where you can configure it.
Insights from Luna
At the bottom of the page there’s a card with insights from Luna, built on your real data: which agent is the most active and what share of runs it makes, whether there are any errors, which enabled agents are sitting idle, and how activity has changed over the week. The “Discuss with Luna” button opens a chat where you can dig into the numbers and change something right away.