Question 1

What is Databricks MCP?

Accepted Answer

It's an MCP (Model Context Protocol) server that connects Databricks to your AI, giving the agent tools to act on your data, through a single remote link.

Question 2

Which AI clients does it work with?

Accepted Answer

Any MCP-compatible client: Claude (web, desktop and mobile), Cursor, ChatGPT with connectors, OpenClaw and other agents.

Question 3

Do I need to code?

Accepted Answer

No. Paste the remote URL into your client, authenticate once, and you're done. No packages to install, no code to write.

Question 4

Is it secure? Can the AI change my data?

Accepted Answer

Access is authorized by you and can be revoked anytime. Permissions follow what Databricks exposes; you approve them when connecting.

Question 5

Is my customers' data stored on mcp.ai?

Accepted Answer

No. Whatever Databricks returns passes through our infrastructure solely to reach the AI assistant you chose, and that content is not stored. By design we don't persist tool-call payloads: the log keeps metadata only (which tool, timestamp, status, duration, size and a hash of the arguments). Your credentials are stored encrypted and you can revoke the connection at any time. Retention periods and details are in the Privacy Policy at mcp.ai/privacy.

Question 6

Is my data used to train AI models?

Accepted Answer

No. Nothing that flows through the connector is used to train or improve AI models, nor shared with third parties for that purpose. The conversation itself, meaning your prompts and the AI's answers, is processed by the AI provider you choose (Claude, ChatGPT and the like), under their own privacy policy.

Question 7

How much does it cost?

Accepted Answer

You start free and upgrade when you need more usage. Billing and limits are managed in your mcp.ai account.

Question 8

Can I cancel anytime?

Accepted Answer

Yes. Plans are monthly and can be cancelled at any time, with no lock-in.

Databricks

Tools (11)

API & agent docs

Guides and tutorials

Como oferecer um MCP dentro do seu produto

Cada usuário do seu app com o MCP dele

Frequently asked questions

Example prompts

Related tags

Databricks

Tools (11)

databricks_cancel_statementCancela um ou mais statements em execução por id. Aceita lista (`statement_ids`).

databricks_current_userIdentifica o usuário do PAT no workspace (whoami via SCIM Me). Útil pra confirmar qual conta/host está conectado e validar o token.

databricks_get_statementStatus + resultado de um ou mais statements por id (polling de queries longas que voltaram PENDING/RUNNING do run_sql). Aceita lista (`statement_ids`).

databricks_get_tableDetalha uma ou mais tabelas (colunas, tipos) por nome completo `catalog.schema.table`. Aceita lista (`full_names`).

databricks_get_warehouseDetalha um ou mais SQL warehouses por id. Aceita lista (`ids`).

databricks_list_accountsLista os workspaces Databricks conectados a este install — host, label.

databricks_list_catalogsLista os catálogos do Unity Catalog visíveis ao PAT (name, comment, owner).

databricks_list_schemasLista os schemas (databases) de um catálogo Unity. Informe `catalog_name`.

databricks_list_tablesLista as tabelas de um schema Unity (name, table_type, data_source_format). Informe `catalog_name` e `schema_name`.

databricks_list_warehousesLista os SQL warehouses do workspace (id, name, state, cluster_size, warehouse_type). Use o `id` em databricks_run_sql (ou deixe o run_sql escolher um RUNNING automaticamente).

databricks_run_sqlExecuta uma instrução SQL num SQL warehouse (Statement Execution API). Retorna colunas + linhas quando termina dentro do wait_timeout; senão devolve statement_id + state pra polling via databricks_get