Elasticsearch Toolbox and elasticsearch-gui are two of the small, single-purpose tools that grew up around Elasticsearch 1.x and 2.x. Elasticsearch Toolbox was a Chrome app: point it at a cluster URL and send queries from the browser. elasticsearch-gui was a plugin you installed into the Elasticsearch folder and opened on port 9200, giving a search page, a query builder and a few graphs. FlatQL is a SQL workbench for Elasticsearch 6 to 9 and OpenSearch 1 to 2 with a hybrid execution engine, exports, query history and team access control. The first two are query runners and browsers. FlatQL is a query engine with a UI around it.
Two small tools#
Elasticsearch Toolbox#
The repository describes it as "a simple, easy to use ide for all you elasticsearch needs right from chrome", and the same description adds that it is "deprecated as chrome apps are deprecated and app was meant for es 2" (GitHub). The repository is MIT-licensed, was archived by its owner on 30 March 2019, and was last pushed in May 2017 (GitHub API). The Chrome Web Store listing still exists at version 0.51, last updated 19 October 2015, with about 4,000 users (Chrome Web Store).
elasticsearch-gui#
Jettro Coenradie's elasticsearch-gui is "an angularJS client for elasticsearch as a plugin" under the Apache-2.0 licence (GitHub). The README describes three ways to explore an index: a website-style search with keywords, advanced search and facets; a query view where you pick fields, facets, highlighting and limits and see how the query is structured; and graphs of some of the data. It installs with bin/plugin install jettro/elasticsearch-gui and opens at http://localhost:9200/_plugin/gui/index.html, or you can use the author's hosted copy with CORS enabled on your cluster. Development targets Elasticsearch v2.0, with branches for 1.1.x and 0.90.8+. The repository is not archived but was last pushed on 3 October 2017 (GitHub API).
Neither tool is a SQL engine. They send Query DSL, or build it from a form, and show the response.
Comparison#
| Elasticsearch Toolbox | elasticsearch-gui | FlatQL | |
|---|---|---|---|
| Licence | MIT (repo) | Apache-2.0 (repo) | Proprietary, Free and Pro plans |
| Form | Chrome app (store) | Elasticsearch site plugin or hosted copy (README) | Web app; desktop app coming soon |
| Target versions | Elasticsearch 2 (repo) | Elasticsearch 2.0, older branches (README) | Elasticsearch 6 to 9, OpenSearch 1 to 2 |
| Query language | Query DSL | Form-built queries, facets | SQL, plus Cluster SQL mode for x-pack SQL or the OpenSearch SQL plugin |
| Joins | No | No | Yes, across indices |
| Window functions, CTEs | No | No | Yes |
| Charts | Not stated | Graphs of some data (README) | Bar, line, area, scatter, pivot |
| Export | Not stated in the repository | Not stated | CSV, TSV, JSON, NDJSON, Excel, Parquet to 5,000,000 rows |
| Access control | None; direct cluster credentials | None; direct cluster credentials | Roles with read or write per index pattern |
| Query history | Not stated | Not stated | Per user and connection, with duration and rows |
| Status | Archived March 2019 | Last push October 2017 | Maintained; see the changelog |
| Price | Free | Free | Free plan; Pro $199 one-time (pricing) |
Where Elasticsearch Toolbox or elasticsearch-gui is the better fit#
- You run an Elasticsearch 1.x or 2.x cluster that cannot be upgraded, and you want a tool built for that generation. FlatQL supports Elasticsearch 6 and later.
- You want something tiny with no account and no backend. Both tools talk from the browser straight to the cluster.
- You want to read the source and keep a copy. Both are permissively licensed and small.
- You want to see how a faceted search request is built. elasticsearch-gui's query view shows the structure as you change the options.
Where FlatQL is the better fit#
- The cluster is on a version from the last several years. FlatQL covers Elasticsearch 6 to 9 and OpenSearch 1 to 2, including Elastic Cloud and Amazon OpenSearch Service.
- You think in SQL. A
GROUP BYonlogs-*becomes a composite aggregation or a date histogram on the cluster, and Show as DSL displays the request next to the result. See Show as DSL and your first query. - You need more than one index in a statement. JOINs, CTEs,
UNION,HAVING,QUALIFYand window functions run in the embedded engine on rows the cluster has already filtered. - You need the result in a file. Exports stream to disk, so a 5,000,000-row CSV does not sit in memory.
- More than one person uses the tool. Credentials live encrypted on the server, members never see them, and the audit log records what ran.
- You want a record of past queries. History and usage reports are built in; the Free plan keeps one day, Pro keeps everything.
Here is the kind of statement that takes a few lines in FlatQL and would be several hand-written aggregation requests otherwise:
SELECT DATE_TRUNC('day', "@timestamp") AS day,
service,
COUNT(*) AS requests,
AVG(duration_ms) AS avg_ms
FROM "logs-*"
WHERE level = 'ERROR'
AND "@timestamp" >= '2026-10-01'
GROUP BY day, service
ORDER BY day, requests DESC;
The filter, the date truncation and the grouping are all pushable, so the cluster does the work and FlatQL returns buckets rather than scanning documents.
Using both#
There is not much overlap to manage. If you still have an old cluster alongside newer ones, keep the old tool for that cluster and use FlatQL for the rest. If a colleague is attached to a form-built facet view, it costs nothing to leave it in place while reports and exports move to SQL.
What you should not do is rely on an archived Chrome app for anything that matters. Chrome apps are no longer the platform they were when Elasticsearch Toolbox shipped, and the project says so itself. For current browser-based options, see the elasticsearch-head comparison, the Dejavu comparison and the roundup of Elasticsearch GUI tools.
Note Dates and version numbers above are from the GitHub repositories and the Chrome Web Store as of November 2026.