Definite

A data team that never sleeps.

[Mike Ritchie @thisritchie]

You can now run a complete AI-native analytics platform in your own cloud.

Lakehouse, semantic layer, automations, data apps, and agents.

Single-tenant. Nothing leaves your environment.

@definiteapp now ships on-prem.

[Mike Ritchie @thisritchie]

Adding data sources is the most frustrating part of setting up analytics.

Fortunately, Opus 4.5 is really good at it:

[Mike Ritchie @thisritchie]

we're building our agent (@definiteapp) with claude code.

We use Skills to tell Claude how to use features within our product. This keeps the main CLAUDE md small, while still giving the agent all the details it needs.

e.g. we have a data-lineage skill and claude knows how and when to use it based on context in the app

[Mike Ritchie @thisritchie]

The average company buys Snowflake, then Fivetran, then a BI tool, then spends 6 months connecting them. We spin it all up in 30 seconds.

[Mike Ritchie @thisritchie]

This comment on HN precisely articulates why text-to-sql fails.

You need "codification" (i.e. semantic layer) to use agents in analytics.

Having an LLM write bespoke SQL to answer every question will fail fast.

e.g. if you ask for "revenue by month" against a Snowflake warehouse with hundreds of tables, you are guaranteed to get different answers over multiple attempts.

@definiteapp uses an agent to help you build a semantic layer so you get the same MRR every time you ask about MRR.

[Mike Ritchie @thisritchie]

It shouldn’t take 6 months and 5 vendors to make sense of your data. @definiteapp raised $10M to give you an AI-native data platform in an afternoon.

Why does it normally take so long? Because the modern data stack is split across three heavyweight products:

1. Data Warehouse: a place to store all your data
2. ETL: pipelines to get data into your warehouse
3. BI: a place to build reports & dashboards

Each of these comes from a different vendor, each with its own price tag and complexity. On top of that, you need data engineers and analysts to stitch them together and keep the system running.

Let’s look at one of our customers, a b2b SaaS that raised $20M. Their current tech stack looks like many of our customers:

* Postgres for their application database
* Hubspot for CRM
* Stripe for payments
* A slew of other sources (customer success platform, Google Sheets, and event data)

They wanted the same thing every company wants: a single source of truth for how the business is performing. Traditionally, this is done with dashboards.

Their plan was to hire a few data people, buy a bunch of tools (Snowflake, Tableau and Fivetran) and hope for the best. Six months and a lot of pain later, maybe they’d have a few dashboards.

Fortunately, they tried Definite one morning and by that afternoon their key metrics were live.

How do we do it?

Definite ships a full data stack (warehouse + pipelines + metrics & reports) in a single app. We’re built on the best open source infrastructure in the market with an AI agent that’s eager to make sense of your data. Instead of hiring, buying, stitching, and waiting; you have the answers you need in minutes running on a data stack that scales. And it's just a chat away.

A huge thank you to our partners at @costanoavc (@JohnCowgill) and @AcrewCapital (@asadkhaliq), to our incredible team, and to our early design partners who have helped shape the product.

And we’re hiring. If you’re a high-agency builder who wants to help reinvent analytics for thousands of operators, let’s talk —> [email protected]

[Mike Ritchie @thisritchie]

We just open-sourced our @meltanodata target for Ducklake (from @duckdb).

* Type conversion is automatic (timestamps stay timestamps).

* Append or merge: choose at runtime.

* Storage is portable (S3, GCS, or local).

* Works with Postgres, MySQL, SQLite, or DuckDB catalogs.

* Timestamp and categorical partitions are built in.

We're already running it in production here at @definiteapp.

[Mike Ritchie @thisritchie]

It's a massive pain to be "data driven". You need:

* a place to store all your data (datalake)
* pipelines to get data into the datalake (ETL)
* a workspace to munge data and share analysis (BI)
* a data team to set all this up and answer questions

Skip all of this. Get @definiteapp.

1. Let's start by adding Stripe:

[Mike Ritchie @thisritchie]

We start you on third base at @definiteapp.

You get a base data model for all our most popular data sources (e.g. @stripe , @HubSpot, @attio, Quickbooks, etc.).

For example, our Stripe model covers MRR, churn, expansion revenue, and lifetime value. You can use all these metrics instantly after you connect Stripe.

If you want to change the way discounts are handled, it's as easy as updating the SQL.

[Mike Ritchie @thisritchie]

@rauchg @definiteapp is v0 for data.

spin up a data lake, pipelines (e.g. sync your stripe data) and dashboards all via our agent.

[Mike Ritchie @thisritchie]

@nearcyan @nabeelqu @definiteapp is cursor for data (backed with the best open source data infra like @ApacheIceberg and @duckdb)

[Mike Ritchie @thisritchie]

our post on @duckdb + @ApacheArrow made the front page

[Mike Ritchie @thisritchie]

@definiteapp in a 🥜

[Mike Ritchie @thisritchie]

our demos at @definiteapp are a little longer than this, but the vibe is identical

[Mike Ritchie @thisritchie]

We (@definiteapp) have a PR open to add predicate pushdown to the DuckDB Iceberg extension.

This is a 0 to 1 change for many production @ApacheIceberg lakes. @duckdb currently downloads every parquet file for every table referenced. This is... not ideal. Most queries would just bomb.

With predicate pushdown, we see:

* simple queries: from several seconds --> subsecond
* complicated queries: From OOM errors --> query actually executes

[Definite @definiteapp]

Choosing the right tools can transform how you handle data. Our latest blog post breaks down the best data analytics tools to help you make smarter decisions.

Check it out: https://t.co/FBar3dXCgk.

#DataAnalytics #TechTools

[Mike Ritchie @thisritchie]

i guess we need light mode to sell to enterprise. fine, here it is ---> https://t.co/RJ9wW66gP0

[Mike Ritchie @thisritchie]

⚫ Definite 2.0

This is the biggest release since we started Definite and I’m insanely proud of it.

We’ve completely redesigned the product with meticulous attention to every detail. We’ve added scalable storage to @duckdb using @ApacheIceberg. And our AI Assistant (”Fi”) is now ready to help you get shit done.

It was painstaking and took months to get right, but the result is worth every minute. Incredibly proud of the team here for pulling this off.

[Definite @definiteapp]

@cereal_learner @thisritchie @stevowang @duckdb we'll do that!

[Mike Ritchie @thisritchie]

Good read here on how and why @NotionHQ built a data lake.

tldr: they saved over $1M by moving off Fivetran + Snowflake. They also set up CDC (change data capture), so ingestion times dropped from days to hours.

If you read this and think "wow, that sounds great, but we'd never build this ourselves", check out @definiteapp (spoiler: we built it for you).