Dex

Create dependable SQL data pipelines more quickly using an AI copilot.

Free Trialdexdata.aiJun 5, 2026

Unified Data Pipelines

Build end‑to‑end workflows covering ingestion, transformation, and orchestration in one place using SQL and Python, then run everything directly on your existing cloud data warehouse.

AI Copilot for Data Work

A context‑aware assistant that helps write SQL, design data models, suggest improvements, and generate boilerplate code based on your existing codebase and metadata.

Software Engineering Guardrails

Native version control (Git‑based), modular code organization, automated testing, and CI/CD pipelines designed specifically for analytics projects—not just application code.

Governance & Observability

A centralized view of data lineage, quality checks, fine‑grained access control, and cost monitoring across all pipelines and environments.

Connector & Collection Layer

Managed connectors and change‑data‑capture pipelines that reliably sync data from a wide range of operational systems into your analytics storage, reducing manual integration effort.

Documentation & Discovery

Automated generation of pipeline and model documentation, making it easier for teams to understand, reuse, and trust data assets across the organization.

What is Dex?

Dex is a unified data engineering platform that combines ingestion, transformation, orchestration, documentation, and governance into a single environment that sits on top of a company’s existing cloud data warehouse. It is built for data and analytics teams that want software-engineering style guardrails for analytics, AI/ML workloads, and process automation, with an integrated AI copilot to speed up SQL and pipeline development. The product focuses on delivering a collaborative, code-centric workbench where teams can build, test, and monitor analytics pipelines using SQL and Python, all while keeping execution on the customer’s own warehouse infrastructure.

Core Features

Unified Data Pipelines: Build end‑to‑end workflows covering ingestion, transformation, and orchestration in one place using SQL and Python, then run everything directly on your existing cloud data warehouse.
AI Copilot for Data Work: A context‑aware assistant that helps write SQL, design data models, suggest improvements, and generate boilerplate code based on your existing codebase and metadata.
Software Engineering Guardrails: Native version control (Git‑based), modular code organization, automated testing, and CI/CD pipelines designed specifically for analytics projects—not just application code.
Governance & Observability: A centralized view of data lineage, quality checks, fine‑grained access control, and cost monitoring across all pipelines and environments.
Connector & Collection Layer: Managed connectors and change‑data‑capture pipelines that reliably sync data from a wide range of operational systems into your analytics storage, reducing manual integration effort.
Documentation & Discovery: Automated generation of pipeline and model documentation, making it easier for teams to understand, reuse, and trust data assets across the organization.

Use Cases & Considerations

Use Cases

Data Engineering Teams Building Central Pipelines: Use Dex as the control plane to ingest from source systems, transform raw data into clean models, and orchestrate complex dependencies—all governed from a single environment.
Analytics Engineers Creating Shared Metrics Layers: Develop and test semantic models and dashboard‑ready tables with version‑controlled SQL, while the AI copilot accelerates writing and refactoring.
Machine Learning & MLOps Teams Managing Feature Pipelines: Curate training datasets and feature stores from the same governed data platform, ensuring consistency between BI, analytics, and ML workloads.
Operations & Finance Teams Automating Reporting: Schedule and monitor KPI‑tracking pipelines, set up alerts for data freshness issues, and reduce manual Excel‑based workflows.
Boutique Data Consultancies Standardizing Client Work: Rapidly deploy a consistent, software‑engineering‑grade data stack for different clients, leveraging Dex’s built‑in guardrails and AI assistance.
Early‑Stage Startups Adopting “Big‑Company” Practices: Implement robust data engineering patterns from day one without hiring a full platform team, using Dex as an accelerator while building on their own warehouse.

Limitations & Considerations

Platform Shutdown: The public Dex service is being shut down following the team’s move to Nubank; new users cannot sign up, and existing public‑cloud tenants face eventual termination.
Migration Lock‑In Risk: Pipelines, models, and orchestration logic are expressed in Dex’s own patterns, meaning that exiting requires re‑implementing everything in another tool—a significant effort if not planned for in advance.
Requires a Modern Cloud Warehouse: Dex is designed to work on top of cloud data warehouses; organizations still relying on legacy on‑premises databases or spreadsheets will not benefit from the platform.
AI Copilot Context Limits: While the AI assistant is context‑aware, its suggestions are based on patterns from the available codebase and may not capture broader business logic or complex edge cases without careful review.
Limited Support and Resources: With the team now focused elsewhere, community support, documentation updates, and bug fixes are minimal or unavailable, making it unsuitable for production‑critical use.
No Self‑Hosted Option: Dex was offered only as a SaaS control plane, which means you cannot stand up a private, internal instance after the public service ends.

How to use Dex

Connect Your Data Warehouse: Start by linking Dex to your existing cloud data warehouse (e.g., Snowflake, BigQuery, Redshift) so that all transformations and orchestration execute there.
Set Up Connectors and Ingestion: Use the managed connector layer or create custom pipelines to sync operational data into your warehouse. Configure change‑data‑capture for real‑time or incremental loads.
Build Transformations and Models: Write SQL and Python data models directly in the built‑in IDE. The AI copilot can generate code, suggest improvements, and auto‑complete based on your project’s context.
Orchestrate and Schedule Workflows: Define dependencies between ingestion, transformation, and downstream tasks. Set schedules and trigger pipelines from the unified orchestration layer.
Apply Software Engineering Guardrails: Commit all code to Git, create modular components, write unit tests for data models, and set up CI/CD pipelines so that changes are reviewed and tested before reaching production.
Monitor and Govern: Use the centralized lineage, quality, and cost dashboards to track pipeline health, enforce access controls, and ensure data reliability across teams.

Pricing & Plans

Historically, Dex sold via enterprise contracts with custom pricing based on scale, workloads, and support needs. A free trial was available to help teams evaluate the platform before committing to a paid plan. Specific list prices were never publicly detailed; engagement typically started with a sales conversation and a proof‑of‑value pilot. However, the team behind Dex has joined Nubank and the public platform is scheduled to shut down, so new commercial agreements are no longer available. For any residual or internal use, pricing information is no longer maintained publicly. Disclaimer: Pricing details may not be current. For the latest status, check the official Dex website.

Platforms

Web Application: A browser‑based IDE and dashboard where users manage pipelines, explore lineage, and interact with the AI copilot.
API Access: Programmatic interfaces for integrating pipeline management and monitoring into custom toolchains or external CI/CD systems.
CLI (Limited Availability): In some deployments, a command‑line interface was offered for power users to trigger jobs and manage configurations from terminals.
Cloud‑Native Execution: Dex itself is a SaaS control plane, but all data processing runs inside the customer’s own cloud data warehouse—no data leaves the customer’s environment for transformation.

Tips & Best Practices

Treat Analytics Projects as Codebases: Leverage Dex’s Git integration to version every model and pipeline. Peer‑review changes through pull requests, just like application code.
Combine the AI Copilot with Human Review: Let Dex’s assistant draft complex SQL or boilerplate, but always validate the logic with domain‑specific business rules and test outputs.
Modularize for Reuse: Break transformations into composable, well‑documented components that different teams can discover and reuse, reducing duplication.
Embed Data Quality Checks Early: Define assertions and quality tests inside the development workflow so issues are caught before they hit downstream BI or ML pipelines.
Monitor Cost in Parallel with Performance: Use the built‑in cost observability to keep an eye on warehouse spend as you scale; small changes in pipeline design can yield significant savings.
Plan for Exit Strategy from Day One: Because Dex’s future as a standalone platform is uncertain, keep pipeline logic as standard SQL/Python where possible, and maintain a mental mapping of how each component would translate to alternative tools.

Who is Dex for?

Data Engineering Teams looking for a single control plane to handle ingestion, transformation, and orchestration across business domains.
Analytics Engineers and BI Developers who want a collaborative, code‑first environment for building and testing shared metrics layers and dimensional models.
Machine Learning and MLOps Engineers that need governed feature pipelines and training datasets originating from the same curated data assets as BI.
Operations and Finance Analytics Specialists aiming to automate complex reporting workflows and KPI monitoring without deep platform engineering support.
Data Consultancies seeking a repeatable, high‑quality stack to implement for clients, reducing time‑to‑value and enforcing engineering best practices.
Startups that want to move fast with a modern, governed data setup but lack the resources to stitch together and maintain separate tools for ETL, orchestration, and testing.

Alternatives

View all

dbt (with Airflow/Prefect)

Combine dbt’s transformation‑focused IDE and testing framework with a dedicated orchestrator for a code‑first, warehouse‑native workflow, though you will need to manage multiple tools.

Coalesce

A GUI‑driven data transformation platform with strong Git and CI/CD capabilities, aimed at analytics engineers who prefer a visual interface over pure code.

Datafold

Focuses on data observability and testing, providing automated lineage, diffing, and quality checks that can be layered on top of any existing transformation stack.

Meltano

An open‑source, code‑centric alternative that can handle ingestion, transformation, and orchestration, though it may require more manual setup and lacks a unified AI assistant.

Atlan

A collaboration and governance platform that adds a metadata layer on top of your existing ETL/ELT tools, helpful for discovering data assets and lineage, but does not execute pipelines itself.

Airbyte + dbt + Airflow (self‑assembled stack)

Build a similar architecture using separate open‑source or managed services, which offers maximum flexibility but shifts integration and maintenance burden onto the data team.

FAQ

Q1. Can I still create a new Dex account?

No. The public Dex platform is being wound down, and new account creation has been disabled. The team has moved to Nubank, and the product is no longer offered commercially.

Q2. What will happen to my existing pipelines after the shutdown?

Public cloud tenants will eventually lose access to the Dex control plane. You will need to export your code and recreate your pipelines using alternative tools. It is advisable to start planning an exit strategy as soon as possible.

Q3. How did the AI copilot work inside Dex?

The AI copilot understood the context of your data models, dialects, and existing codebase. It could generate SQL snippets, suggest model improvements, explain complex queries, and even help refactor pipelines—all within the integrated IDE, accelerating development for both experts and less‑technical analysts.

Q4. Does Dex work with any data warehouse?

Dex was designed to be warehouse‑agnostic, with support for major cloud data warehouses like Snowflake, Google BigQuery, Amazon Redshift, and Databricks. All transformation and orchestration workloads ran directly on the connected warehouse; Dex itself never extracted or stored your data.

Q5. Is there any way to keep using Dex after the shutdown?

As a standalone public platform, no. The underlying technology might live on internally at Nubank, but external users will not have access. For those who admired Dex’s approach, the best path is to adopt the same design principles (code‑centric pipelines, Git‑native workflows, embedded testing) using today’s self‑assembled or alternative unified platforms.