SOLUTIONS

PRICING

TEAM

BLOG

DOCS

INTRODUCING TOWER

Ship data flows on managed lakehouse infrastructure faster

Tower unifies serverless data flow orchestration and Iceberg lakehouse management into a single control plane, so your engineers can focus on delivering insights to your business and fresh data to your AI.

Technical Docs

solutions

From data pipelines to data agents

Orchestrate Python-native pipelines, dbt workflows, notebooks, agents, and tools.
All in one platform.

Tower gives you a reliable, open lakehouse built on Apache Iceberg and compatible with Snowflake, Spark, and what comes next.

Run Data Pipelines

Run extraction, loading, and transformation pipelines securely and reliably on Tower.

Learn More

Build dbt Models

Develop and run dbt Core models directly on Tower, without relying on additional managed services.

Learn More

Deploy Data Agents

Power data agents with fresh, reliable business data and ship faster with rapid iteration on Tower.

Learn More

benefits

Why Tower?

Get a Python-native orchestrator of data flows and optionally use a reliable, open lakehouse built on Apache Iceberg and compatible with Snowflake, Spark, and what comes next.

Data Flows

Data Lakehouses

Fully-managed data flow orchestration

Flows: Choose between Pythonic control flows and Agentic workflows

Observability: Gain full flow observability, logs, alerts and metrics

Configuration: Manage your settings and secrets in dedicated environments from dev to prod

Open data integration, ETL, and AI Inference

Use your favorite ETL Tools: Use community-developed libraries like dltHub, Polars, Apache Data Fusion

Unlock the power of agents: Integrate with OpenAI, LangChain, Ollama, llama.cpp and HuggingFace

Lakehouses managed and optimized by Tower

Iceberg REST Catalog: Use our managed service or bring your own catalog

Table maintenance: Improve performance, clean up wasted storage

Ingestion: Stream data into your lakehouse or load it in batches

dbt Core

1 import tower
2 
3 def main() -> None:
4   commands = tower.dbt.parse_command_plan()
5   workflow = tower.dbt(
6     project_path=project_path,
7     profile_payload=tower.dbt.load_profile_from_env(),
8     commands=commands,
9   )
10   workflow.run()

dbt Core

dltHub

DuckDB

Lakekeeper

Build using open-source tools

Data Transformations: Use dbt Core, dltHub, Polars and DuckDB to build the layers of your lakehouse

Catalogs: Lakekeeper, Apache Polaris, Snowflake Open Catalog, Cloudflare R2

dbt Core

1 import tower
2 
3 def main() -> None:
4   commands = tower.dbt.parse_command_plan()
5   workflow = tower.dbt(
6     project_path=project_path,
7     profile_payload=tower.dbt.load_profile_from_env(),
8     commands=commands,
9   )
10   workflow.run()

dbt Core

dltHub

DuckDB

Lakekeeper

Build using open-source tools

Data Transformations: Use dbt Core, dltHub, Polars and DuckDB to build the layers of your lakehouse

Catalogs: Lakekeeper, Apache Polaris, Snowflake Open Catalog, Cloudflare R2

dbt Core

1 import tower
2 
3 def main() -> None:
4   commands = tower.dbt.parse_command_plan()
5   workflow = tower.dbt(
6     project_path=project_path,
7     profile_payload=tower.dbt.load_profile_from_env(),
8     commands=commands,
9   )
10   workflow.run()

dbt Core

dltHub

DuckDB

Lakekeeper

Build using open-source tools

Data Transformations: Use dbt Core, dltHub, Polars and DuckDB to build the layers of your lakehouse

Catalogs: Lakekeeper, Apache Polaris, Snowflake Open Catalog, Cloudflare R2

TOWER VS COMPETITION

How we compare

Tower combines data flow orchestration, flexible compute options, and analytical storage management into a single platform.

GitHub Actions

Airflow + Self-Managed Compute

Dagster

Tower

Flow Orchestrator

Defining data or control flows

Partial (1)

Yes

Flexible Compute Options

Self-hosted Runners for Data Security + Serverless Runners in Tower Cloud

Yes

Partial (2)

Partial (3)

Yes

Analytical Storage

Management tools for Lakehouses

Yes

Multi-Tenant Platform

Control Plane APIs to run apps and manage users

Partial (4)

Yes

GH Action has some orchestration primitives but no flow support

Requires use of a Managed Airflow Service

Dagster Self-Hosted version requires hosting both Data and Control planes

APIs for triggering runs are supported; User management APIs depend on Dagster Self-Hosted vs Dagster Cloud

MULTI-TENANT PLATFORM

Designed for Builders

Whether you are a data team building an internal platform, a data + AI consultancy, or a SaaS provider, Tower's multi-tenant management platform gives you the right building blocks.

Adaptive and Powerful APIs

Save months building your platform

Best in class infrastructure

Brought to you by engineers who built Snowflake, Databricks, Google Cloud Dataflow, Puppet, and more

Hardened APIs

Built-in user management, isolated environments, and seamless Python app deployment and execution

Headless Data Stack

Add Tower’s open lakehouse storage and multi-engine capability to your data platform

Flexible Data Plane

Choose between self-hosted runners or serverless infra

Self-Hosted Runners

Process sensitive data in your own cloud account or run local AI inference on data that can’t leave your premises

Serverless Infrastructure

Use our fully-managed service deployed to reliable cloud providers

Observability Single Pane-of-Glass

Unified logs, metrics, alerts, scheduling and control flows that work across environments

AI-Ready Data Infrastructure

Build the future-proof data infra for your AI

MCP server

Tower MCP server guides your AI Coding Assistants in deploying apps and launching runs

Agent Tools

Turn Tower apps into data agent tools and query your data lakehouse or onboard external data

TESTIMONIALS

What our users and partners say

Inflow supports customers with sensitive data workloads that must remain on-premises, leveraging a diverse ecosystem of data tools, including Sling Data, dbt, Pentaho, and others, to choose the best fit for each use case. Tower’s self-hosted runners allowed us to consolidate a subset of our customers into a single managed environment, providing a true single pane of glass for observability and a simple, reliable way to deploy and update pipelines.

Itamar Steinberg

CEO, Inflow Systems Ltd

Itamar Steinberg

CEO, Inflow Systems Ltd

Itamar Steinberg

CEO, Inflow Systems Ltd

Docker for data? If you go to a docker conference you will hear "portability", "fast onboarding", "lower cognitive load", "implicit access", "microservices", "decentralisation". Who does this for data? Check out dlthub.com/blog/tower

Adrian Brudaru

dltHub Co-founder

Adrian Brudaru

dltHub Co-founder

Adrian Brudaru

dltHub Co-founder

I have long suffered at the hands of both the Docker and Python ecosystems. Tower solves both. Plus, it plays nicely with scheduling, orchestration, and open table formats! If you're working with Python, you should try it.

Chris Riccomini

Apache Airflow PMC

Chris Riccomini

Apache Airflow PMC

Chris Riccomini

Apache Airflow PMC

dltHub and Tower partner on delivering data movement and transformation. Tower’s data compute infrastructure is tightly integrated with dltHub’s pipelines and notebooks. Users get an end-to-end solution and don’t have to waste time on integrations.

Matthaus Krzykowski

Co-Founder & CEO, dltHub

Matthaus Krzykowski

Co-Founder & CEO, dltHub

Matthaus Krzykowski

Co-Founder & CEO, dltHub

In the age of ChatGPT, it’s no longer the lack of Python or SQL knowledge that’s holding people back from working with data, it’s the sheer pain of setting up local environments for data access. Tower’s customizable and portable Python runtime is a true game changer for data driven organizations. You can stop fixing secrets and dependencies and start creating value from data in no time.

Simon Rosenberger

Enterprise AI Lead, Taktile

Simon Rosenberger

Enterprise AI Lead, Taktile

Simon Rosenberger

Enterprise AI Lead, Taktile

Migrating to Tower didn’t just streamline our infrastructure — it gave us a competitive edge. With lower costs, faster pipelines, and higher forecast accuracy, a-Gnostics can deliver even more value to customers in energy markets where every percentage point matters.

Andy Starzhinsky

VP, a-Gnostics

Andy Starzhinsky

VP, a-Gnostics

Andy Starzhinsky

VP, a-Gnostics

tower blog

New features and milestones

Tower Self-Hosted Runners for Sensitive Data Workloads

Some data workloads are too sensitive, costly, or require strict data sovereignty, so they can’t run in public clouds.

Build a better open lakehouse with Tower and Lakekeeper

When we evaluated Iceberg REST catalog options, Lakekeeper immediately stood out.

Introducing Webhooks: Real-Time Run Events from Tower

Imagine if data pipelines could detect new data as soon as it arrived, rather than checking for it on a fixed schedule.

Why data teams should pick data lakehouse architectures

Why you should choose a lakehouse for your business

Takeaways from PyData Boston and Python Summit Warsaw

Earlier in December, we gave two talks and shared the Tower story at the booths of two Python data events in Boston and Warsaw.

Simpler Tower onboarding with the Tower Iceberg Catalog and our example apps

Launching Tower Catalog, a REST catalog for Apache Iceberg tables that is fully managed by Tower.

Tower Self-Hosted Runners for Sensitive Data Workloads

Some data workloads are too sensitive, costly, or require strict data sovereignty, so they can’t run in public clouds.

Introducing Webhooks: Real-Time Run Events from Tower

Imagine if data pipelines could detect new data as soon as it arrived, rather than checking for it on a fixed schedule.

Takeaways from PyData Boston and Python Summit Warsaw

Earlier in December, we gave two talks and shared the Tower story at the booths of two Python data events in Boston and Warsaw.

Build a better open lakehouse with Tower and Lakekeeper

When we evaluated Iceberg REST catalog options, Lakekeeper immediately stood out.

Why data teams should pick data lakehouse architectures

Why you should choose a lakehouse for your business

Simpler Tower onboarding with the Tower Iceberg Catalog and our example apps

Launching Tower Catalog, a REST catalog for Apache Iceberg tables that is fully managed by Tower.

Tower Self-Hosted Runners for Sensitive Data Workloads

Some data workloads are too sensitive, costly, or require strict data sovereignty, so they can’t run in public clouds.

Takeaways from PyData Boston and Python Summit Warsaw

Earlier in December, we gave two talks and shared the Tower story at the booths of two Python data events in Boston and Warsaw.

Why data teams should pick data lakehouse architectures

Why you should choose a lakehouse for your business

Introducing Webhooks: Real-Time Run Events from Tower

Imagine if data pipelines could detect new data as soon as it arrived, rather than checking for it on a fixed schedule.

Build a better open lakehouse with Tower and Lakekeeper

When we evaluated Iceberg REST catalog options, Lakekeeper immediately stood out.

Simpler Tower onboarding with the Tower Iceberg Catalog and our example apps

Launching Tower Catalog, a REST catalog for Apache Iceberg tables that is fully managed by Tower.

View Blog

Power Your Team with Tower

Get a Python-native orchestrator of data flows and optionally use a reliable, open lakehouse built on Apache Iceberg and compatible with Snowflake, Spark, and what comes next.

Tower gives you a reliable, open lakehouse built on Apache Iceberg and compatible with Snowflake, Spark, and what comes next.

1	import tower
2
3	def main() -> None:
4	commands = tower.dbt.parse_command_plan()
5	workflow = tower.dbt(
6	project_path=project_path,
7	profile_payload=tower.dbt.load_profile_from_env(),
8	commands=commands,
9	)
10	workflow.run()