OpenSearch Dashboards Alternative
For searching and editing documents day to day, a desktop client such as DynoTable is the alternative to OpenSearch Dashboards. Dashboards stays where you build visualizations and dashboards, run reports and manage Index State Management (ISM) policies. For reading data it gives you Discover, the Dev Tools console and Query Workbench.
DynoTable opens each index as a tab, next to your DynamoDB tables, and signs every request with the AWS profile you already use. You search with Lucene or filter pills, edit a document and review the diff before it commits, and run SQL that joins an index to a DynamoDB table.
Amazon OpenSearch Service domains, OpenSearch Serverless collections and a self-hosted endpoint all open the same way. There is nothing to install on the domain.
When to use OpenSearch Dashboards and when to use DynoTable
| Your workflow | Choose |
|---|---|
| Build a visualization, a dashboard, a report or an alert | OpenSearch Dashboards |
| Manage ISM policies, index templates or security roles | OpenSearch Dashboards |
| Move indexes between hot, UltraWarm and cold storage | OpenSearch Dashboards |
| Search an index and read the matching documents | Either |
| Edit or bulk-change documents and review the diff before it lands | DynoTable |
JOIN an index to a DynamoDB table, or GROUP BY across both | DynoTable |
| Open indices beside DynamoDB tables, switching AWS profiles | DynoTable |
Dashboards answers "what does this index look like over time". DynoTable answers "which documents are these, and what do I change". Each Dashboards install belongs to one domain. AWS's newer OpenSearch UI spans several domains and collections from the console, and it is still a browser app with no DynamoDB in it.
Sign in with your AWS profile, not a Dashboards login
AWS states the constraint plainly: "Dashboards does not natively support IAM users and roles" (AWS docs). A domain's Dashboards needs one of SAML, Amazon Cognito, fine-grained access control with a username and password, or an IP-based access policy, often behind a proxy. Someone has to set that up per domain before you can open it.
DynoTable signs each request with your profile's AWS credentials, the same
ones the AWS CLI reads from ~/.aws. It finds your domains and Serverless
collections with those credentials too, so a new domain shows up in the
⌘K Indices group without a login page. The identity still needs
data access: an access policy on a domain, a data access policy on a
collection.
A VPC domain is the one case where neither tool gets a shortcut. AWS says "your computer must be able to connect to the VPC", usually through a VPN, transit gateway or proxy (AWS docs). DynoTable talks to the domain endpoint directly, so it needs the same network path Dashboards does.
A self-hosted cluster connects through an endpoint URL you set per profile. That connection sends no AWS signature and no username or password, so it fits a development cluster without the security plugin, and it also works on an offline profile next to DynamoDB Local. A cluster that requires credentials needs a different client.
Search an index without creating an index pattern
Discover starts from an index pattern you pick from a dropdown, searches with DQL by default or with Lucene, and "displays only the documents that are included in the time filter's time interval" (OpenSearch docs). That model suits log data with a timestamp. It is a detour for a product catalog or an orders index you want to look up by ID.
In DynoTable you open the index itself from ⌘K. The search box takes Lucene and completes field names from the mapping and values from the index as you type. Filter pills compile to query DSL. When the index has a date field you get a time range with presets from 15 minutes to a year, and when it has none, there is no time filter in the way.
For anything the pills can't express, switch the tab to raw mode and write a JSON query DSL body or a Lucene line. The editor lints both against the mapping, and every query you run lands in that index's history. Column sorts run on the server, so a sort orders the whole index rather than the rows already loaded. A free-text search highlights the matching words in each cell.
Edit a document and review the change before it lands
Discover documents viewing, filtering and export. It does not document an editor. In Dashboards you change a document by writing the REST request in Dev Tools, and it is applied the moment you send it.
DynoTable opens a document in an editor with three views: Plain (the JSON), Document (a form, easier for long text) and Flow (a read-only graph). ⌘S stages the change, and Pending changes shows a per-field diff you commit in full or in part; ⌘⇧S stages and commits in one step. Every commit is journaled locally, so you can revert it later from History.
On a domain, the update carries the version of the document you read. If someone changed the document since, the commit stops with a conflict you can rebase instead of overwriting their write. Serverless collections expose no version check, so a collection tab shows a Last-write-wins badge and the later write replaces the earlier one.
Bulk edit and bulk delete work over a selection or every document a query matches, staged for review. For changes too large to stage, a direct run writes without staging. It runs as a native by-query task on a domain and is not revertible, and the confirm dialog says so before it starts.
Query Workbench joins, and a JOIN across two engines
Query Workbench runs SQL and PPL against your domain and can download the result. It is read-only by design: it "does not support delete or update operations through SQL or PPL" (OpenSearch docs).
Its joins are where the SQL plugin is weakest. OpenSearch's own limitations
page says "JOIN queries are supported on a best-effort basis", that "the
JOIN query does not support aggregations on the joined result", that join
queries "automatically terminate after 60 seconds", and that a query with
aggregation and join "does not support pagination"
(OpenSearch docs).
We ran a join with a GROUP BY through the SQL plugin on OpenSearch 2.19,
against an orders index of 150 documents and a customers index:
SELECT c.country, COUNT(*)
FROM orders o JOIN customers c ON o.customerId = c.customerId
GROUP BY c.countryIt answered with HTTP 200 and no error. The result had 150 rows instead of 6,
one per order, and every count was null. The schema came back with three
columns, o.COUNT(*), c.country and c.COUNT(*), for a two-column query.
The same join without GROUP BY returned correct rows, so the failure only
shows once you aggregate, and nothing in the response tells you it happened.
DynoTable's SQL Workbench runs the join in the app instead of handing it to the SQL plugin. One statement can join an index to a DynamoDB table, or to another index, including one on a different domain:
SELECT c.country, COUNT(*) AS orders, SUM(o.amount) AS revenue
FROM dynamodb.orders o
JOIN opensearch.customers c ON o.customerId = c.customerId
GROUP BY c.countryAn index joins on a keyword, number or date field, and GROUP BY works over
the joined rows. The conditions OpenSearch can answer travel with each side's
search, so fewer documents come back, and the full WHERE still runs on the
joined rows. The trade-off is that the join runs on your machine, so a join
over millions of rows is bounded by what the app can fetch and hold.
OpenSearch SQL covers the join rules for each engine.
What stays in Dashboards
DynoTable has no charts, visualizations or dashboards. It also does not manage index templates, ISM policies, alerting monitors, reports, notebooks or security roles. Those are Dashboards' job and stay there.
What DynoTable does manage is the index itself. Create an index from ⌘K (it takes its mapping from the first document you write; a hand-written mapping is still a REST call), change replicas and the refresh interval on a domain, delete an index behind a typed confirmation, and export its definition as a REST request or Terraform. Serverless collections manage shards and the refresh interval themselves, so those settings are not offered there.
What it costs
Dashboards comes with every domain at no extra line item, but "Dashboards runs on the hot data nodes in the domain" (AWS docs), so heavy Dashboards use shares capacity with your searches.
DynoTable runs on your machine. Its Free plan has no time limit and needs a
free DynoTable account. Searching an OpenSearch index is part of it, along with
browsing DynamoDB tables, PartiQL SELECT, exports and MCP reads. A paid seat
adds editing and staging, the SQL Workbench and Smart Tables over tables and
indices, and the AI agent. Individual seats start at $9/mo billed annually
($108/yr); see pricing for monthly and Team rates.
Try DynoTable against the same indices
Download DynoTable for macOS (Apple Silicon or Intel), Windows or Linux, turn on OpenSearch for your profile, and open an index from ⌘K. The trial is free for 30 days, no credit card required. The OpenSearch docs walk through connecting a domain, a Serverless collection or a self-hosted cluster.
FAQ
Is there an alternative to OpenSearch Dashboards for browsing and editing data?
Yes. DynoTable is a desktop client for Amazon OpenSearch Service domains, OpenSearch Serverless collections and self-hosted clusters. It searches an index with Lucene or filter pills, stages document edits as a reviewable diff, and runs SQL that joins an index to another index or to a DynamoDB table. It has no visualizations or dashboards, so Dashboards stays for those.
Does DynoTable replace OpenSearch Dashboards?
Not entirely. DynoTable covers searching, reading and editing documents, bulk changes, index creation and SQL across engines. Visualizations, dashboards, reports, alerting, ISM policies, index templates and security roles stay in OpenSearch Dashboards.
Can I log in to an OpenSearch domain with my AWS profile instead of Cognito or SAML?
In DynoTable, yes. It signs every request with your AWS profile's credentials, and the identity needs data access through the domain's access policy or the collection's data access policy. OpenSearch Dashboards itself does not natively support IAM users and roles, so it needs SAML, Cognito, fine-grained access control or an IP-based policy.
Can OpenSearch SQL join two indices?
On a best-effort basis. OpenSearch's documentation says the JOIN query does not support aggregations on the joined result, terminates after 60 seconds by default, and does not support pagination when combined with aggregation. DynoTable's SQL Workbench runs the join in the app, so GROUP BY works over joined rows, and one side can be a DynamoDB table.
Related
- Browse the full comparison hub for every DynoTable alternative.
- See also the AWS DynamoDB console comparison for the DynamoDB half of the same AWS account.
- Read how DynoTable connects to OpenSearch.
Last verified 2026-10-29. OpenSearch is a trademark of LF Projects, LLC. AWS, Amazon OpenSearch Service and DynamoDB are trademarks of Amazon Web Services; referenced here for identification only.