Main Sequence Documentation
Main Sequence is a platform for building data products, deploying workflows, and exposing them as reusable platform resources.
The Main Sequence platform allows you to:
- rapidly build and deploy data products and data workflows as a unified API with normalized table contracts through
MetaTables andDataNodes - rapidly deploy RBAC-enabled dashboards on the platform
- rapidly deploy agents using the Google Agent SDK
The key idea is that you can focus on development and deployment, while the platform handles the DevOps layer.
Choose a reading path
Tutorial
Use the tutorial if you want a guided path from project setup to canonical tables, DataNode update workflows, orchestration, and dashboards.
The main tutorial path is intentionally CLI-only.
Recommended starting points:
- Part 1 — Setting a Project (CLI)
- Part 2 — Working With MetaTables
- Part 3 — Creating a Data Node
- Part 3.1 — Role-Based Access Control
- Part 3.2 — Create Your First API
- Part 6 — Turn Your Project Into an Agent
Knowledge
Use the Knowledge section if you already know the workflow you are building and want the deeper conceptual guide.
Key entry points:
- Data Nodes
- Command Center
- Command Center Widget Data Contracts
- MetaTables
- Constants and Secrets
- Scheduling Jobs
- Streamlit Dashboards
CLI
Use the CLI section for command-focused operational work:
Reference
Use the generated reference when you need the API surface directly:
What this SDK covers
The SDK is organized around a few main areas:
mainsequence.meta_tables:DataNodes, persistence, orchestration, and update workflowsmainsequence.client: API client models for projects, jobs, tables, platform resources, and sharingmainsequence.cli: themainsequencecommand-line interface
If you are new here
The fastest onboarding path is usually:
- create a project
- model a canonical table with a backend-managed
MetaTable - build your first
DataNodeas an opinionated MetaTable-backed update workflow - understand RBAC and shared resources
- add an API or another consumer surface
- schedule jobs
- expose the result through dashboards, downstream readers, or agent-facing project surfaces
That sequence is exactly how the tutorial is structured.