How Do You Search Your Old LinkedIn Posts?

LinkedIn gives you no way to search your own post history. Flux fixes this with semantic search: open the Explore page or call the search_posts MCP tool, type a natural-language query like "that carousel about pricing," and Flux returns your most relevant past posts ranked by meaning. Because it matches on meaning and not just exact words, it finds the right post even when your query doesn't share a single keyword with it.

What you'll learn

Why LinkedIn's Activity feed makes finding old posts nearly impossible
How to search your archive by meaning with the Explore page
How to run the same search programmatically with the search_posts MCP tool
How to find posts similar to an example with find_similar_posts
Practical workflows: repurposing, prospect research, and avoiding repetition
How Flux's hybrid search works under the hood

Why can't you search your own LinkedIn posts?

LinkedIn has no search for your own post history. The only way to revisit what you've written is to open your Activity feed and scroll, post by post, in reverse-chronological order. There's no filter by topic, no keyword box, and no way to jump to "the thing I posted about hiring last spring."

That's fine for your last few posts and useless for everything older. If you posted about pricing strategy eight months ago and want to find it, you're stuck scrolling past hundreds of posts hoping to recognize it. Flux replaces that scroll with a search box that understands what your posts are about.

How do you search your archive by meaning?

The Explore page (/explore) gives you a single search box that takes natural-language queries and returns your most semantically similar past posts, ranked by similarity. Instead of guessing the exact words you used, you describe what the post was about and Flux surfaces the matches.

Search "that carousel about pricing" and Flux finds the post even if the word "pricing" never appears in it — a post titled "How we landed on our plans" is a strong match by meaning. This is the difference between semantic search and a plain keyword find: you search for the idea, not the phrasing.

Explore uses hybrid search, combining semantic matching (based on embeddings) with lexical matching (keyword overlap). That means an exact keyword like a product name still ranks where you'd expect, while conceptual queries still work when you can't remember the wording. The Explore page is gated by the explore_enabled flag and an Edge-tier subscription.

How do you search your posts programmatically?

The search_posts MCP tool runs the same semantic search engine as the Explore page, available programmatically. You pass it a natural-language query and get back your ranked, most-relevant posts — the same hybrid ranking, just callable from an AI workflow instead of the browser.

This is what makes search useful inside a drafting loop. Before writing a new post, you can ask your assistant to "find my posts about onboarding so I can reference them," and search_posts returns the relevant ones to pull quotes, stats, or angles from. The tool does the retrieval; your draft builds on what you've already said well.

How do you find posts similar to one you already have?

The find_similar_posts MCP tool takes a post URL or ID and returns posts with similar content from your archive. Where search_posts starts from a text query, find_similar_posts starts from an existing post and asks "what else have I written like this?"

This is the tool to reach for when you're developing a theme. If you're writing about onboarding, point find_similar_posts at your current draft's closest published post and it surfaces everything adjacent you've written before. It often turns up posts you'd forgotten — patterns in your own thinking that are hard to spot by scrolling.

What can you do with post search?

Search turns your archive from a write-only log into a reference library. A few workflows that pay off immediately:

Content repurposing: Search for old high-performing posts on a topic, then refresh and repost the best ones instead of starting from a blank page.
Prospect research: Search a tracked profile's library before a meeting to find the posts that show what they care about right now.
Avoiding repetition: Before publishing, search your own library to confirm you haven't already covered the same ground recently.
Finding reference material: When someone asks "didn't you write about X?", find the post in seconds instead of scrolling for ten minutes.

How does Flux search work under the hood?

Flux embeds every post using the same model that powers topic clustering, so the meaning of each post is captured as a vector and stored per-post in Supabase. When you run a query, the search-posts Edge Function embeds your query the same way and finds the posts whose meaning sits closest to it.

The ranking is hybrid: semantic similarity finds conceptually related posts, while lexical matching ensures exact keyword hits aren't drowned out by their semantic neighbors. A search for a specific product name still surfaces the posts that name it, even when a dozen posts are conceptually nearby. You get the recall of meaning-based search without losing the precision of a keyword match.

Frequently asked questions

Can I search my LinkedIn posts by keyword, or only by meaning?

Both. Flux uses hybrid search, so exact keywords like a product or person's name rank where you'd expect, while conceptual queries still work when you can't remember the exact wording. You don't have to choose between keyword and semantic search — every query uses both.

Does Flux search by impressions or views?

No. Flux tracks engagement (likes, comments, and shares), not impressions, and search ranks results by relevance to your query, not by performance. If you want your best-performing posts on a topic, search for the topic and then sort by engagement in the results.

Can I search someone else's posts, not just my own?

Yes, for profiles you track. Searching a tracked profile's library is how prospect research works — you find the posts that reveal what someone is focused on before you reach out or meet.

What's the difference between search_posts and find_similar_posts?

search_posts starts from a text query and returns your most relevant posts. find_similar_posts starts from an existing post (by URL or ID) and returns posts like it. Use the first when you know what to search for, and the second when you have an example and want more of the same.

Why can't I see the Explore page?

The Explore page is gated by the explore_enabled flag and an Edge-tier subscription. If you're on Edge and still don't see it, the feature may not be enabled on your account yet — the search_posts MCP tool runs the same search engine in the meantime.

On this page