Bluesky Analytics: Read Your Post Stats in ChatGPT

Bluesky analytics can tell you which posts prompted a response. They can't tell you how many people silently read them. That distinction matters when you ask an AI assistant what worked: a polished explanation is only useful if the numbers support it.

With Postory connected, you can ask ChatGPT or Claude to read your imported Bluesky archive, summarize engagement and find posts worth revisiting. Start with the available counts, check how fresh they are, and compare posts that had a fair chance to collect responses.

What Bluesky analytics can measure

Postory imports four engagement counts for your own Bluesky posts and replies:

Metric	What it counts	A useful question
Likes	Likes on the post	Which posts received the most likes?
Replies	Replies to the post	Which posts started conversations?
Reposts	Reposts of the post	Which posts were shared onward?
Quotes	Posts quoting yours	Which posts prompted additional commentary?

These are different actions. A reply can disagree with you, and a quote can challenge your point. A larger engagement total doesn't automatically mean a better response. Open the original posts before deciding to repeat an approach.

Bluesky's notification guide also distinguishes replies, quotes, likes and reposts. Your notifications help you inspect the conversations behind the counts; the counts alone don't describe their tone.

In Postory's performance summary, engagement means likes plus replies plus reposts plus quotes. That gives you a consistent way to compare saved posts, but it isn't a count of unique people. One person can take more than one action.

There is no view or impression count in this Postory workflow. Don't ask the assistant to turn engagement into a reach estimate or an engagement rate based on views. A missing denominator is missing information, not zero.

Bring your archive into the conversation

Connect your Bluesky account to Postory, then connect Postory to your AI app. The Bluesky MCP connection guide covers that setup. Once the connection is available, work in this order:

1. Choose the account. Ask the assistant to list your connected accounts and select the intended Bluesky account. Name it explicitly if you manage several.
2. Check the archive. Ask what posts and own replies are imported, the available history, and when the archive was last updated.
3. Import or update the missing content. Wait for the import to finish before requesting the analysis. Bluesky imports and metric refreshes through Postory use no credits.
4. Request a dated summary. Ask for posts first, with your own replies summarized separately. Include the dates, post count and freshness notes in the answer.

A useful opening prompt is:

“Use my Bluesky account. Check the imported history and update it if needed. Summarize posts published in the last 30 days, showing likes, replies, reposts and quotes. Tell me what data is missing.”

The period selects when the posts were published. Their metrics are the counts saved when those posts were last checked; they are not a log of every interaction that happened inside that period. A post published near the start of the month has had longer to collect responses than one published yesterday.

Check freshness before comparing results

An imported number is a snapshot. If you read it again without a metric refresh, the post may have gained more likes or replies on Bluesky in the meantime.

Ask the assistant to show the metric timestamps and any excluded items. Postory's summary distinguishes posts without usable metrics from posts with a measured zero. It also warns when earlier history doesn't cover the comparison period. Don't turn either gap into a claim that your performance fell.

Use this follow-up when the answer looks incomplete:

“Before comparing periods, check whether both are covered by my archive. Refresh the available metrics if they're old, and explain any remaining gaps.”

You can still learn from a partial archive. Label the result “among the imported posts” and avoid claims about your entire account. If the assistant returns a few interesting examples, ask whether it selected them from a complete list or a relevance search. Examples illustrate a pattern; they don't establish how common it is.

Ask questions the numbers can answer

Start with one decision, such as which subject deserves another post. Ask the assistant to show the underlying post links and separate each engagement count, so you can inspect its reasoning.

For a content review:

“Show my strongest Bluesky posts from the selected period by engagement. Include their text, dates, links and separate counts. Suggest recurring subjects, and distinguish observations from guesses.”

For a discussion review:

“Which of those posts received replies? Compare their opening questions and subjects. Don't assume the replies were positive without reading the conversation.”

For finding earlier work:

“Search my imported Bluesky posts for discussions of my newsletter. Show relevant examples and their saved engagement counts.”

That last prompt searches your imported archive. It doesn't search everyone on Bluesky, and a relevance search isn't a complete ranking of everything you've published.

When comparing periods, keep their lengths equal and include the number of posts in each. A higher total alongside more posts answers a different question from higher engagement per post. Also compare similar formats and allow for post age. Treat a recurring pattern as a reason to try another post, not proof that a format caused the result.

Start reviewing Bluesky posts with Postory

Use Postory's AI-app connection to bring your archive and its metrics into the same conversation where you plan your next post. Ask for one dated review, inspect the examples, and choose one idea to develop.

Try Postory free and start with the Bluesky posts you've already published.