Matthew Hefferon

@shuvo_saha__ Did you find the button?

I freaking love this metabase man I spent 10 mins looking for a button to sort the numbers in descending order But then I realized, I know SQL. So I just slapped an ORDER BY clause on the query and done The flexibility is insane

Join the Hackathon buff.ly/VrJKNBa

From my phone, I sent a voice note to Claude. Cowork received it → Dispatch routed it as a task → @metabase MCP queried the data → answer came back to my phone. #MetabaseAI

The official Metabase MCP server is here! 🔹 streamable HTTP transport 🔹 same permissions and access 🔹 OAuth 2.0 authentication 🔹 leans on the semantic layer

AI just went open source.  Welcome to Metabase 60 🫶

@sseraphini @metabase We have a Pinot community connector docs.pinot.apache.org/build-with-pin…

how to integrate Pinot on @metabase ?

Give it a try and provide your feedback in GitHub: github.com/metabase/metab…

We do dope stuff @metabase 🤖🎉

@vincz_a @metabase I'll shoot you a dm to get some more info.

@HeffHQ @metabase Hmm no, still the same issue :/

New in v59: AI SQL generation on self-hosted 🤖 Bring your own Anthropic key, describe what you need, get SQL back. Less time writing queries, more time getting answers.

@ujsquared @metabase Thanks for flagging! I took a look and I’m not seeing that on my end.

@metabase hey metabase team, great update! just this one little bug on your site's footer needs a look.

@vincz_a @metabase I know Claude had an outage today. I just gave it a quick test and it ran successfully on my end. Maybe you ran into that?

@metabase I just tried this out but I got an error "Something went wrong, please try again". Running self hosted 0.59.2 - any idea what is going on? It does not look like the request is going to Anthropic.

@robertocommit @metabase We've tested 6+ other models and Claude has performed best.

@robertocommit @metabase We started with Anthropic but plan to test other models in the near future.

@cdolan92 @metabase 👍

@metabase It is! Nice appreciate the OSS feature!

With Data Studio, teams can: • Transform raw data into analytics-ready structures • Define metrics and dimensions once, use everywhere • Publish datasets others can rely on • See dependencies before making changes See what we built and why → metaba.se/data-studio

A new way to shape your data and define shared meaning as your analytics grows.

@HarelDan @Boris24 Curious to hear what got tricky at scale.

Best free tools I discovered this year for data work: • Evidence.dev — SQL-to-dashboard, no code • Metabase — Self-hosted BI, insanely good • Datasette — Turn any SQLite DB into an API in 30 seconds • DBeaver — Universal database client, handles everything All free. All production-ready.

I’m a fan of labels but too many make trends harder to see. Here’s how I’d clean it up in @metabase

@peer_rich @metabase I dig the open dashboard! If you want another set of eyes, happy to help. DMs open.

weird flex: db so big, @metabase struggling to query the tables 😆 will see what quick fix i can do to bring back some metrics on cal.com/open sadly low priority but on it