Esploro
Esploro 0.10.1

Secure database rails
for your AI agents.

A beautiful, full-featured SQL client for macOS. Browse, edit, and query PostgreSQL, MySQL, and MariaDB with ease.

Connect your agents through a local, authenticated MCP server. Read-only queries. Built-in limits. No database passwords shared.

Download for Apple Silicon Explore MCP server
Or install with Homebrew
brew install --cask matija/tap/esploro

macOS 13 or later · Free for personal use · Download for Intel

For paid work, choose a commercial license.

Esplorocedar.demo
retailorderscustomersproductsinventoryshipments
retail.ordersData
Fictional retail orders
order_refcustomerstatustotal
ORD-2401Juniper Studiopaid84.50
ORD-2402Harbor Supplyshipped129.00
ORD-2403Meadow Goodspending56.75
ORD-2404Birch Workshoppaid215.40
4 of 842 rows24 ms
Illustrative app views. All connections, tables, and data are fictional.

MCP · Model Context Protocol

Your agent writes SQL.
Esploro handles access.

Your agent writes the SQL. Esploro supplies the tools to find saved connections, understand schemas, and run validated, read-only queries. No database passwords to paste into your agent.

  1. Your AI agentAsks questions. Writes SQL.
  2. Esploro MCPAuthenticates. Validates. Runs.
  3. Your databasesPostgreSQL · MySQL · MariaDB
Find your saved databases list_connections
Discover connection names and IDs, without exposing database credentials.
Understand the schema inspect_schema
Inspect tables, columns, types, and relationships before the agent writes a query.
Ask questions of your data execute_query
Run read-only SELECTs, including joins and aggregations. Count orders, compare totals, or find the latest users.
For example, ask your agent
“Which customers spent the most since January 2026?”

The agent queries through Esploro

SELECT c.display_name,
       COUNT(*) AS orders,
       SUM(o.total) AS revenue
FROM retail.customers c
JOIN retail.orders o USING (customer_id)
WHERE o.ordered_at >= DATE '2026-01-01'
  AND o.status IN ('paid', 'shipped')
GROUP BY c.customer_id, c.display_name
HAVING COUNT(*) >= 3
ORDER BY revenue DESC LIMIT 3;
Illustrative query results with fictional customer revenue
display_nameordersrevenue
Juniper Studio184820.50
Harbor Supply153910.00
Meadow Goods122854.75

Illustrative workflow with fictional data, not a live query.

Saved connections. One configuration. Ready to ask.

  1. Open Esploro.Save your database connections and keep the app running.
  2. Configure MCP access.Open Settings → MCP, grant access to your chosen connections, and enable the server. Use its endpoint and token in your agent’s MCP settings.
  3. Ask your first question.Ask the agent to find your databases, inspect a schema, and run a read-only query.
MCP configuration · JSON 
{
  "mcpServers": {
    "esploro": {
      "type": "http",
      "url": "http://127.0.0.1:19482/mcp",
      "headers": {
        "Authorization": "Bearer <YOUR_TOKEN>"
      }
    }
  }
}

Paste this into your MCP client’s configuration. Replace <YOUR_TOKEN> with your token from Settings → MCP. Keep the token private. Client formats vary; use a client that supports Streamable HTTP and bearer headers.

Guardrails built in. Access you can inspect.

Local and authenticated
The server listens only on your Mac’s loopback address. A bearer token stored in Keychain protects access.
Read-only, with limits
SQL validation blocks writes and schema changes. Queries use read-only transactions, with a 30-second timeout and up to 1,000 returned rows.
Visible query activity
Review the latest 20 query attempts in Settings → MCP Activity, including SQL, status, duration, and returned row counts.

The MCP token grants access to permitted connections. New connections start without agent access. Use restricted database roles. Read-only access is not complete isolation. Query results can reach your agent’s provider, including cloud services.

Settings / MCPcedar.demo
MCP serverRead-only

Permitted connections

  • cedar.demoPermitted
  • cedar.stagingPermitted
  • cedar.sandboxNo access
Local endpoint
http://127.0.0.1:19482/mcp
Bearer token
••••••••••••••••
Choose which connections your agents can access.

From the first row
to the next query.

Your daily database tasks belong together. Browse tables, run SQL, save queries, and manage permissions in Esploro.

Start with the data.

Select rows together for bulk actions. Filter by a cell value, copy as JSON or CSV, or duplicate a row.

Use table filters to narrow the results, then copy the generated SQL for your next query.

retail.orderscedar.demo
Fictional order selection
order_refcustomerstatustotal
Selected: ORD-2401Juniper Studiopaid84.50
ORD-2402Harbor Supplyshipped129.00
Selected: ORD-2403Meadow Goodspending56.75
ORD-2404Birch Workshoppaid215.40
2 rows selectedDelete selected
Select several rows and act on them together.

Write SQL. Inspect the results.

Use joins, common table expressions, and window functions to explore your data. Inspect results below the editor, then save useful queries for next time.

Querycedar.demo
WITH customer_sales AS (
  SELECT c.display_name,
         COUNT(*) AS orders,
         SUM(o.total) AS revenue
  FROM retail.customers c
  JOIN retail.orders o USING (customer_id)
  WHERE o.status IN ('paid', 'shipped')
  GROUP BY c.customer_id, c.display_name
  HAVING COUNT(*) >= 3
)
SELECT *, DENSE_RANK() OVER (
  ORDER BY revenue DESC
) AS rank FROM customer_sales
ORDER BY revenue DESC LIMIT 3;
3 rows18 ms
Fictional customer revenue results
display_nameordersrevenuerank
Juniper Studio184820.501
Harbor Supply153910.002
Meadow Goods122854.753
Your query and its results, in one view.

Make room for
your way of working.

Choose a theme, adjust fonts, and set the table row height. Your settings stay in the app.

Settings / Appearancecedar.demo
Appearance
Theme
Tairiki Light ✓
Interface font
Readex Pro
Editor font
SF Mono 12 px
retail.ordersSELECT * FROM retail.orders
WHERE total > 75;
Your theme. Your fonts.

Your database
stays yours.

Esploro connects from your Mac to your database. Your queries and results do not pass through an Esploro cloud service.

Inspect the source on GitHub
Passwords in macOS Keychain
Esploro stores database passwords in the system credential store.
MIT source code
You can read the code, report problems, and contribute changes.
A clear privacy policy
Commercial license checks use an online service. The website uses analytics. Read the privacy policy.

A focused choice for Mac.

Esploro focuses on PostgreSQL and MySQL, including MariaDB. If you need more database engines, compare the options.

Database support and source availability. Sources reviewed September 19, 2026.
FeatureEsploroPostico 2TablePlusDBeaver Community
PostgreSQLYesYesYesYes
MySQL / MariaDBYesNoYesYes
Open sourceYesNoNoYes
Apple Silicon buildYesYesYesYes

Sources reviewed September 19, 2026: Postico, TablePlus, DBeaver.

One app for your SQL.
And your agents.

Free for personal, non-commercial use.
For paid work, a commercial license is required.

Download for Apple SiliconSee commercial licenses

Download for Intel · macOS 13 or later · Release notes

Compare your options.

All articles →