A faster Databricks client than DBeaver.

Three real pain points

The cold start. DBeaver runs on the JVM, and every launch pays a startup tax, several seconds you feel every morning even on a fast Mac. QueryFlow launches like a native app, because it is one.

Catalog navigation as an afterthought. Unity Catalog's three-level structure, catalog, schema, table, works in DBeaver's generic tree, but it's the same tree it uses for every JDBC source, not something built around how Databricks actually organizes data.

No AI panel without add-ons. Asking a question in plain English and getting working SQL back against your real catalog names isn't part of DBeaver's core experience.

Feature comparison

Where DBeaver still wins

DBeaver's connector list is enormous, and it's free and open source. If you connect to a dozen niche databases QueryFlow doesn't support, or you'd rather not pay for a client, DBeaver remains a reasonable choice. This isn't about replacing it everywhere, just about not using it as your daily Databricks driver if speed and a real catalog view matter to you.

A worked example

A query against a gold-layer table, the kind you'd write daily:

SELECT product_category, sum(revenue) AS total_revenue
FROM retail.gold.sales_by_product
WHERE sale_date >= current_date() - interval 7 days
GROUP BY product_category
ORDER BY total_revenue DESC;

Autocomplete in QueryFlow offers retail.gold.sales_by_product and its columns because the explorer already knows the catalog's shape, not because you typed enough characters for a generic guess to land.

Honest migration note

Switching means re-adding your Databricks connection by hand, the hostname, HTTP path and a token or service principal again. There's no automatic import from DBeaver's connection list, and saved queries you want to keep need to be copied over manually.

If you keep both installed

You don't have to pick one forever. Keep DBeaver for the odd niche connector it supports that QueryFlow doesn't, and use QueryFlow for daily Databricks work. They're separate apps pointed at the same warehouse; nothing about using one affects the other.

A word on cost, either way

Neither client changes what a Databricks SQL warehouse costs to run. Compute is billed by warehouse size and uptime, not by which tool submitted the query, so the real cost lever is the warehouse's auto-stop setting and how big you've sized it, not the client you query it from.

Keeping both around during the switch

You don't have to delete DBeaver the day you start QueryFlow. Run both for a week or two against the same warehouse, compare results on a query you already trust, and let the daily habit shift on its own rather than forcing a hard cutover before you're confident the new editor covers what you need.

What actually carries over on day one

Nothing automatic. The Server Hostname, HTTP Path, and either a token or service principal credential are all you need to reconnect. If you've built up a personal library of saved DBeaver scripts, plan on copying over the ones you actually still use rather than migrating everything on principle.

A reasonable order to do it in

Start with the queries you run daily, get those working and confirm the results match what DBeaver gave you, then go back for anything used less often. That gets you productive in the new client fastest, with the rarely-used queries left for whenever you actually need them again.

What the Ask panel can help with here

If a query you're porting over is complex enough that retyping it feels error-prone, paste it into a scratch tab and ask the panel to explain what it does before trusting it's correct in the new environment. That's a sanity check, not a replacement for reading the query yourself.

Where DBeaver's SQL formatter still helps

If DBeaver's formatting conventions are ones your team has standardized on, running an old query through it once before pasting into QueryFlow is a reasonable habit to keep, even after you've otherwise switched clients.