Studio is the web workbench that ships with every ArcadeDB server. Write queries in eight languages, explore results as a graph, manage schema, users and backups, and profile slow queries, all from a browser and at no cost. Add the Studio AI Assistant when you want to ask questions in plain English.
There is nothing to install separately. Start a server, then point a browser at its HTTP port:
for a default local install that is http://localhost:2480. Studio uses the same users as the server.
This command starts ArcadeDB 26.10.1 with the OpenBeer sample database preloaded, the one in the screenshot above:
docker run --rm -p 2480:2480 -p 2424:2424 \
-e JAVA_OPTS="-Darcadedb.server.rootPassword=playwithdata \
-Darcadedb.server.defaultDatabases=Imported[root]{import:https://github.com/ArcadeData/arcadedb-datasets/raw/main/orientdb/OpenBeer.gz}" \
arcadedata/arcadedb:26.10.1Log in as root with the password playwithdata. Running ArcadeDB inside your own JVM instead?
Studio is part of the server, so see client-server mode for how to run one alongside your application.
The Query panel is the everyday workspace. Pick a language, run the command with
Ctrl+Enter / Cmd+Enter, and Studio picks the best view for the result.
SQL, SQL Script (multi-statement), OpenCypher on the native engine, Apache TinkerPop Gremlin, GraphQL, MongoDB, Redis and PromQL for time series. Syntax highlighting follows the selected language.
Query a vector, hybrid or full-text index from a form: pick the index, paste the query vector or a Lucene query, set k and an optional filter. The same operations as the HTTP vector and full-text endpoints.
Table with sorting, paging, CSV export and an inline record editor; Graph; Chart (line, area or bar, from any language); JSON; and Explain with an optional flame graph of per-step timings.
Recent commands reload with one click, saved queries stay a click away, and a searchable reference of SQL functions and methods sits next to the editor. An Auto Limit cap keeps interactive runs safe on large datasets.
The Graph view draws the vertices and edges a query returns and keeps them on the canvas, so you can expand the graph incrementally instead of re-running bigger and bigger queries. It is the fastest way to sanity-check a knowledge graph or follow a suspicious chain of transactions in a fraud detection model.
Beyond queries, Studio is the administration console for the server and every database on it.
ArcadeDB has a built-in MCP server. From the Server panel you enable it, choose which operations it may perform, pick the users allowed to use it, and copy a ready-made configuration for Claude Desktop, Claude Code or Cursor. API tokens for those clients are created in the Security panel.
That is free and works with your own AI client. The step-by-step guide is in ArcadeDB MCP Server: Connect Your LLM to Your Database, and the bigger picture of using the graph as context for an LLM is on the GraphRAG page.
The AI Assistant is a paid add-on that turns Studio into a chat-driven workspace. Ask about your database in plain language and get answers, schema insights and executable queries back. It reads your schema, can run read-only queries on your behalf, and proposes commands you run with one click.
Buy it standalone, or get it with any paid support plan. Every plan includes a monthly AI usage allowance per server or license. Allowances are pooled across your servers (three Silver servers give you $150 of AI use a month), start over on the first of each month, and do not roll over. When the allowance is used up, the assistant pauses until the next month; you can upgrade at any time.
Prices in USD. Support plans can also be billed yearly. Plan details, checkout and terms are on the pricing page.
After purchase you receive a subscription key. Open the AI tab in Studio, enter the key and accept the AI Assistant Terms of Use. The subscription is stored on the ArcadeDB server and shared by every Studio user connected to it.
Start a server with the command above and open http://localhost:2480. When you want the AI Assistant, pick a plan on the pricing page.
Yes. Studio ships with every ArcadeDB server and is part of the Apache 2.0 open-source project. Only the AI Assistant panel is a paid add-on; every other Studio feature is free.
Start an ArcadeDB server and point a browser at its HTTP port. For a default local install that is http://localhost:2480. Studio uses the same users as the server, so you log in with your ArcadeDB credentials.
The Query panel runs SQL, SQL Script, OpenCypher, Apache TinkerPop Gremlin, GraphQL, MongoDB, Redis and PromQL, and has forms for vector, hybrid and full-text index search. Results can be shown as a table, a graph, a chart, JSON, or an execution plan.
The AI Assistant is available standalone at $19.99 per month per license ($199.99 billed yearly), with $20 of AI use per month. It is also included in every paid support plan, with a monthly AI allowance per server of $50 for Silver, $100 for Gold and $250 for Platinum. Allowances are pooled across your servers, start over on the first of each month and do not roll over.
No. Write operations (inserts, updates, deletes, schema changes) are never executed automatically; they are returned as command blocks for you to review and run. In Review First mode the assistant does not run anything at all, and in Auto Run Read Only Queries mode it may only run read-only queries, each shown in the conversation.
To answer, the assistant sends schema definitions, metadata, query structures and small data excerpts to a third-party LLM provider. Per the AI Assistant Terms of Use, your data is never used to train any model and is not retained by the provider beyond the request. For sensitive workloads, use test or anonymized data and a read-only Studio user.