Dejavu 3.0: The missing Web UI for Elasticsearch

Dejavu 3.0 — the missing web UI for Elasticsearch is here 🎉 🎉! It’s been an amazing journey thus far: Since our first release in 2015, we have crossed a lifetime total of 475,000 Docker pulls, have over 11K active Chrome extension installations, and over 5,100+ stars 🌟 on our Github repository.

With this major release, we’re adding a rad feature that allows building and visually testing search UI in minutes. It’s also a complete re-write of the code to improve performance by 2x-3x and reduce Dejavu’s memory footprint by over 50%. You can read about the full changelog here.

While building appbase.io — a hosted search service over the past 3 years, one of the pain points we continually kept coming across from our users was how to build the right search experience. It turns out that there is a steep learning curve associated with understanding how a search engine like Elasticsearch works.

In a nutshell, building a search experience today requires being able to import data without breaking a sweat, tweak mappings and the data itself as you are testing stuff, and being able to tune the queries related to search.

Dejavu is an open-source MIT licensed Web UI for Elasticsearch. It demystifies the experience of building and testing a search UI. Dejavu started out as a rich data browser for Elasticsearch and with the 3.0.0 release, it allows you to build and test the search UI for your Elasticsearch index.

What can you do with Dejavu 3.0?

• Import JSON or CSV data to your Elasticsearch index,
• Get a rich data browsing experience (hide/show data fields, filter data, query views, and CRUD operations),
• Connect to multiple indexes and browse nested JSON data as columns, including an ability to apply faceted filters on the data,
• Build a search UI and test relevancy of your index in a matter of minutes,
• Write search queries with a GUI composer.

Additionally, for appbase.io users, Dejavu can do a few more things:

• Edit mappings on the fly (e.g. change an int field to a text field and then reindex your entire data),
• Create multiple profiles for testing search UI views (save every search setting you update, and build search for multiple views or A/B test an individual view).