Ein Thema was mich seit Steemit bewegt und ich auch für Newbies oft höre - eine Basis Search Funktionlität wäre super - lest weiter, hab grad Zeit zu schauen und manchmal zu meckern :-)!

What if you need a former post for a job application but no way to find it?

I think this is one of the major things since Steemit days never was considered for improvement, right? Tagging @holoz0r - given I said I do not wanna complain or ask for certain things haha!

In the early days I was trying to keep track on posts in series / tags by adding links in all new posts to find them easier - i then ignored / forgot it - I do regret this though. Why? On Hive you have no real chance to have good search for your own (and others) content, no filters no nothing - only chance you scroll down 99 years of activity to find certain posts - can we change that? Sometimes you are lcky to ask for a tag and in tredning / recent you might find it faster via scrolling.

A basic filter search feature - is that doable?

Asking @blocktrades - sorry mate - is that something so difficult to implement for the chain or do we need a dev team / third party to solve this? Maybe not major issue for Hive these days but I think this is a comodity thing usually for a platform - just hinting on it politley with a friendly smile. Would be even happy with a filter combination of three filters:

• Tags (of post)
• Date (search between two time points)
• Author name

An Example - searching for tag #RockClassics (Step 1)

I created that one and think did most posts around it (over 60). When I search on Peakd, no matter if i do with # or without same results - @jarvie just an example - Hive.blog no better - I see this:

Tag correct, first 4 entries by a cool guy (3 of them from Steemit days) - how is that structured in which order you see results?

Example Step 2

I then can go on Peakd for "recent" and get this - a bit better but still wrong or some tells me in which order this is displayed (for sure not recent) based on tag (frist tag, second tag, third tag)? As recent it could not be the date of publishing as this would mean my recent posts about "RockClassics in 2025" fell off the search engine or from the chain :-).

Example Step 3

You can select also "rewards" which should show the most rewarded post - straight forward - not important for me (results hmmmm - not true - not showing em)

Example Step 4

You can even select searching posts keyword based on "relevance" - see the result below. So what is "relevance" when I search for "RockClassics" in teh search bar? Still no idea what is shown here, what are the criteria to show what? REsults for relevance:

Rant over - was just a small suggestion / ask for feature (my clients do this everyday and I know from some newbies here that started and struggled in finding content they are interested in)

---

However I asked ChatGPT about this

Here is what AI came up with, maybe something to consider what they gave me in the following??

---

Why the internal search feels bad

1. Very basic search matching

   • Hive’s blockchain search relies mostly on matching exact words in titles/tags rather than full-text relevance across body content.
   • This means you miss posts that use synonyms or related phrases.

2. No ranking by relevance

   • Results are often just chronological or by trending score — not by how closely they match your query.

3. Lack of advanced filters

   • You can’t easily narrow by:

     • Author
     • Date range
     • Community/tag
     • Popularity (views, votes, comments)
     • Content type (blog, video, image-heavy)

4. Poor handling of multi-word queries

   • A search like "strategic content marketing" might bring up posts with any of those words, in any order, with no prioritization.

5. No synonym / fuzzy matching

   • Searching AI won’t necessarily bring up posts tagged artificial-intelligence.

---

Filters that would fix this

Reveal spoiler

Example search interface

Search: [ strategic content marketing ]
Filters:
  Author: [ uwe-lang ] 
  Tags: [ blockchain, marketing ]
  Date: [ Last 30 days ]
Sort by: [ Relevance ▼ ]

---

Technical approach

• Hive already has an API that can return posts by tag, author, and date — but it’s fragmented.
• Peakd could add a server-side search index (e.g., Elasticsearch or Meilisearch) to crawl Hive content and provide:

• Full-text search
• Relevance scoring
• Faceted filters

This would make searching on Hive frontends feel more like searching on Medium or Reddit, rather than scrolling endless “trending” pages.

---