9.1sec
to plan and push down a 48-model project against the live warehouse
Daspire Transform compiles version-controlled SQL models against the raw tables you already landed — incremental, tested, and lineage-aware. Build staging, marts, and a metrics layer in the warehouse where your analysts already live. dbt-compatible. No orchestration glue. No stale dashboards.
Transform isn't a notebook and it isn't a black box. It's the modeling layer that lives in your warehouse: version-controlled SQL, materialized the cheapest correct way, tested on every run, with lineage parsed straight from the code. Here's what that buys you in practice.
Models are just SQL files with ref() and source(). Daspire resolves dependencies, builds the DAG, and compiles to native warehouse SQL — no rows ever leave Snowflake or BigQuery.
Declare uniqueness, not-null, accepted-values, and referential tests in YAML — or write custom assertions in SQL. Daspire runs them on every build and halts the downstream DAG the moment one fails.
Daspire parses the SQL — not your comments — to build a column-level graph. Click net_revenue and see every model, join, and source it flows through, plus what breaks if you drop it.
Pick view, table, or incremental merge per model. On a 40M-row fact table, an incremental run rewrites the last hour — not the last two years. Daspire generates the merge key and the predicate for you.
Every Daspire model pack is a version-controlled set of staging models, marts, and tests tuned to a commerce source — maintained in lockstep with the connectors that feed them. Fork it, override it, or run it as-is. The lineage stays intact either way.
The logic that defines "revenue" shouldn't live in a BI tool's saved query, a scheduled notebook, and three analysts' heads. Daspire Transform makes it one version-controlled, tested, lineage-aware model — compiled where the data already is.
Write models in the SQL you already know, declare tests and metrics in YAML, or compose the graph in the visual builder. Every path compiles to the same lineage-aware DAG and runs against your warehouse.
Transform treats every model as versioned, tested, deployable code — with the CI gates and audit trail a regulated team expects.
Slim builds compile only changed models and their children — green check before merge, not after.
When a model slows down or a test flips red, you see the run, the diff, and the blast radius downstream.
Tag a source column as PII once; Daspire propagates the policy to every model that touches it.
Reproduce any table at any point in time from the run log. Auditors reconcile to the commit.
If yours isn't here, ask in chat — we don't gatekeep technical conversations behind a sales call.
Daspire Transform is dbt-compatible — point it at your existing dbt project and it runs. What you add on top: managed compile + run infrastructure, column-level lineage parsed automatically, a built-in metrics layer, and CI that runs the affected DAG on every pull request. You keep your SQL; you stop maintaining the orchestration around it.
No. Every model compiles to native SQL and runs as pushdown inside Snowflake, BigQuery, Databricks, or your warehouse of choice. Daspire orchestrates the run and reads metadata — the rows never move.
On an incremental model, Daspire generates the merge predicate so a run rewrites only rows newer than the last build — typically minutes of warehouse compute, not a nightly full rebuild. You pick view, table, or incremental per model; switch any time.
Daspire parses the compiled SQL on every build to derive column-level lineage — there's no separate graph to maintain. If a model's SQL changes, its lineage updates on the next run, and impacted downstream models are flagged before they break.
Yes. Models are git files reviewed in pull requests, so analysts contribute SQL while engineers own the framework, tests, and CI. The visual builder lets less-SQL-fluent contributors compose models that compile to the same DAG.
No — Transform runs against any raw tables already in your warehouse, however they got there. It pairs naturally with Extract & Load (the raw tables arrive lineage-ready), but it's not a requirement.
14-day free trial — no credit card. Point Daspire at your warehouse, fork a commerce model pack, and run your first tested build in minutes.