Introducing Mako

We built Mako because every SQL client we tried fell into one of two buckets: powerful but bloated, or lightweight but dumb. The AI-powered ones were worse — chat wrappers that hallucinate column names and generate queries against tables that don't exist.

Mako is different. It's a SQL client where the AI is the interface.

Not Another Chat Wrapper

Most "AI SQL tools" paste your question into a prompt and hope for the best. Mako actually inspects your schema before generating anything. It knows your tables, your columns, your relationships. When it writes a query, every identifier is real.

This means you can ask in plain English:

-- "Show me users who haven't logged in for 30 days"
SELECT * FROM users
WHERE last_login < NOW() - INTERVAL '30 days'
ORDER BY last_login DESC;

And the query runs. No fixing column names. No guessing at table structures.

Nine Database Drivers, One Interface

Mako connects to the databases you actually use:

• PostgreSQL — the workhorse
• MySQL — because legacy systems exist
• MongoDB — aggregation pipelines included
• ClickHouse — for your analytics warehouse
• BigQuery — Google's serverless data warehouse
• SQLite — lightweight and local
• MSSQL — enterprise environments
• Cloudflare D1 — edge SQL
• Cloud SQL (Postgres) — GCP-managed databases

Same interface, same AI, same experience — regardless of what's underneath.

Schema-Aware AI That Doesn't Hallucinate

Here's what makes Mako's AI actually useful: it reads your database schema before answering. Every suggestion references real tables and columns. If you ask about revenue, it looks for revenue, amount, total, or whatever your schema actually calls it.

The AI also builds context over time. It learns which tables you query most, what joins make sense in your schema, and how your naming conventions work. We call this self-directive memory — the more you use Mako, the better it gets at understanding your database.

Built for Teams

Mako is workspace-based. Your team shares database connections (encrypted at rest), saves queries to a shared library, and collaborates in real time. No more Slack threads with SQL snippets. No more "which database has the users table?"

Every query is saved, searchable, and synced across devices.

Open Source, Free Forever

Mako is MIT licensed and free to use. No credit card, no trial period, no feature gates. Run it in the cloud at app.mako.ai or self-host it from the GitHub repo.

We believe the best developer tools are open. Mako is no exception.

Try It Now

Head to app.mako.ai, connect a database, and ask your first question. The AI takes it from there.