Maia product overview

An introduction to Maia and its three core components: Maia Foundation, Maia AI Agents, and the Context Engine.

is made up of three integrated components: , , and . provides the cloud-native execution infrastructure for building and running data pipelines. operate within that infrastructure to build, maintain, and optimize pipelines using natural language. supplies the institutional knowledge that reason against, so that the pipelines they produce are consistent with your data standards and business logic.

These docs use “Maia” to refer to the platform as a whole. Where it’s necessary to distinguish between the three components, the full component name is used.

---

Maia Foundation

is the cloud-native execution layer that underpins . It provides the infrastructure for connecting to data sources, building and running pipelines, and governing data operations across your organization.

Designer

is the visual canvas where you build and manage data pipelines in . Pipelines are composed of connected components that you configure to load, transform, and orchestrate data. uses a Git-based branching model so that multiple team members can work on pipelines simultaneously and promote changes through environments in a controlled way.

For more information, read Using .

Pipelines

supports three pipeline types:

• Orchestration pipelines define the sequence and control flow of data loading tasks, including conditional branching, iteration, and error handling.
• Transformation pipelines reshape and prepare data within your warehouse using drag-and-drop components, without requiring you to write SQL.
• Test pipelines run another pipeline and check its results, so you can verify that pipelines behave correctly before deploying them to production.

can build all three pipeline types when you describe what you need in or the chat panel.

To learn about pipelines and the components you can use to build them, read Building pipelines and Components overview. For information about test pipelines specifically, read Test pipelines.

Connectors

includes over 150 connectors for loading data from sources into your cloud data platform. You can also create custom connectors to integrate any REST API that returns a JSON response.

For more information, read Connectors overview and Custom connector overview.

Streaming

supports near-real-time data collection through streaming pipelines. These pipelines capture and synchronize data changes from source databases and write them to a target data warehouse.

For more information, read Streaming pipelines.

Runners

A executes the tasks defined in your pipelines. In a Full SaaS deployment, Matillion hosts and manages the on your behalf. In a Hybrid SaaS deployment, you host the in your own cloud environment (AWS, Azure, Google Cloud, or Snowflake). A handles execution of streaming pipelines.

For more information, read overview.

Deployment and security

can be deployed as a fully Matillion-hosted SaaS solution, as a Hybrid SaaS deployment where the runs in your own cloud, or natively inside Snowflake. All deployment models use a pushdown architecture, meaning data is processed within your cloud data platform and does not pass through Matillion infrastructure.

supports Single Sign-On (SSO), Multi-Factor Authentication (MFA), and Role-Based Access Control (RBAC). For compliance details, read Cloud platform security and compliance.

---

Maia AI Agents

are the agentic layer of . They operate inside the execution environment, interacting with and your data platform directly on your behalf. You can communicate with using natural language in the chat panels in and .

can build orchestration and transformation pipelines, create custom connectors, query and explore your data warehouse, manage project files, analyze and optimize existing pipelines, and commit and push changes to Git. provides the knowledge base that draw on during these tasks, so that the pipelines they build reflect your organization’s data standards and business rules.

are enabled at account level by default, meaning all users in your organization can prompt them immediately. Admins can disable or restrict their ability to sample data from the account settings.

To learn more, read overview.

Mission Control

is the organizational center for . Use it to create and manage autonomous tasks that work through without requiring you to stay in the chat panel. When creating a task, you select a knowledge graph from to give the context they need for that task.

You can also create and manage tasks programmatically using the agent tasks API. For more information, read Mission Control and Using the agent tasks API.

---

Context Engine

Public preview

is the knowledge layer of . It stores and structures institutional knowledge—including business definitions, compliance rules, data lineage, table and column relationships, and warehouse metadata—so that can reason against it when building and maintaining pipelines.

Knowledge is organized into knowledge graphs, each populated by crawlers that connect to your data warehouse and pipeline execution history on a schedule. also learns from your chat history with ; each conversation can automatically update semantic information in the knowledge graph, so the knowledge base evolves as your data landscape does.

When you create a task in or work with in the chat panel, you can select a knowledge graph to use, giving the context they need to produce results that are consistent with your data standards.

For more information, read Context Engine.