Tool documentation

How to read Bluesky post analytics

Understand recent Bluesky post statistics, the interactions-per-post formula, sample dates and exclusions, with a practical review method and public-data limits.

Review your recent Bluesky posts

Open profile analytics, enter a public handle, DID or profile link, and run the analysis. No account connection or app password is needed. The report is a quick review of a bounded public sample, useful for finding posts worth revisiting or choosing examples for a bot’s documentation.

First read the number of included posts, the examined entry count and the sample dates. A sample covering three days and one covering three months answer different questions, even when each contains the same number of posts. Coverage describes whether the requested feed ended, not whether every post in the account’s lifetime is available.

What enters the sample

The tool requests at most four pages of 50 author-feed entries, with an eight-second reading budget after the profile check. It includes unique, visible posts authored by the requested DID, excluding replies and repost feed entries. A repost of the author’s own older post is still excluded. Hidden content is not a shortcut to extra statistics.

Deleted, unavailable or filtered content can be absent. When a later page fails, the report marks its coverage as limited and keeps the usable sample. If the initial feed cannot be read, it shows an error rather than an empty success.

Understand the calculations

The four totals add the included posts’ current likes, replies, reposts and quotes. Interactions per included post equals the sum of these four totals divided by the number of included posts; an empty sample has a zero average. The top ten posts use the same sum for ordering.

For example, 20 included posts with 80 likes, 10 replies, 8 reposts and 2 quotes have 100 observed interactions, or 5 per post. These are current counters on the sampled posts, not interactions received only between the displayed post dates. They are not unique people, impressions or a measured engagement rate.

Make a useful comparison

Write down the capture date, included count and date window before comparing two reports. Review the actual top posts: a single exceptional publication can shift the average. Do not compare the result directly with a service that includes replies or uses a different period and denominator. For the text and context of one post, open the post viewer.

Use the API or MCP

The tools API exposes GET /api/v1/tools/profile-analytics?q=HANDLE; MCP exposes analyse_profile_posts. Both return the same sample counts, dates, coverage, totals and top ten posts. One call costs three quota units. Successful results share a five-minute cache across interfaces, so repeatedly refreshing cannot produce a real-time dashboard.

Privacy and limits

The tool only reads public data and cannot publish or modify a profile. It stores no follower-growth history and does not estimate reach, views or audience demographics. Inputs are not logged as analytics properties, and hidden identities are omitted. Read the privacy policy for platform logging details. Treat the report as a reproducible description of one recent public sample, with its limits kept alongside the numbers.

Profile analytics

Understand your recent public posts.