Skip to main content
Product

Introducing the Sigma CLI: Manage Sigma from your terminal or coding agent

TJ Wells
TJ WellsProduct Advocate
Jonathan Cook
Jonathan CookProgrammatic Services
August 20, 2026
6 min read

Today we're announcing the Sigma CLI. The Sigma CLI (command-line interface) puts Sigma's entire public REST API in your terminal.

To use it, install the CLI, run a single login command, and then you can read and change anything in Sigma that our API exposes. This includes workbooks, data models, connections, teams, permissions, schedules, and more. Just as importantly, it gives AI coding agents like Claude Code, Codex, Cursor and CoCo a structured way to build in Sigma on your behalf, without first teaching them how our API works.

The CLI is now available to every Sigma customer on macOS, Linux, and Windows.

Why did we build the Sigma CLI?

We built the Sigma CLI because the way people build is changing. Whether technical or non-technical, most people now start their work with an AI assistant and then move into another platform. Non-technical users want to prompt their way into instant dashboards and applications, and data teams increasingly build through scripts, pipelines, and coding agents rather than only through a browser.

This means that teams today usually start in one of two places:

  1. A business-facing AI interface like Claude, ChatGPT, Snowflake Cortex, or Databricks Genie
  2. A developer environment like Claude Code, Codex, Snowflake CoWork, or the Databricks CLI

Business users can explore artifacts in familiar AI experiences, while developers can build more directly in a code-first environment.

Sigma supports both patterns. With Sigma, you can build applications directly from the terminal (the workbook spec API is currently in a private beta and coming soon to public beta; request access through your Customer Success Manager). Whether someone begins by asking a question in natural language or building from the terminal, that work can flow into Sigma to become a dashboard, workflow, or full application on live data.

The real advantage is portability. No matter which AI interface a team chooses, the Sigma CLI provides a consistent path from prompt to production. Teams can change tools without rebuilding workflows or starting over. The interface may vary, but the destination stays the same: useful, production-ready Sigma AI applications, delivered faster.

What can I do with Sigma CLI?

The Sigma CLI brings Sigma workflows into the terminal, giving developers and administrators a consistent way to work with workbooks, data models, and other Sigma resources. You can use it interactively for day-to-day administration, automate it through scripts and CI pipelines, or connect it to coding agents that build and manage Sigma content on your behalf. Here are three common ways to get started:

Terminal output listing Sigma workbooks using the Sigma CLI
Run a single CLI command to list every Sigma workbook in your org.

1. Admin and bulk-operation workflows

The Sigma CLI is the strategic home for admin and bulk operations. This includes managing workbooks, data models, and other resources at scale.

2. Scripted/CI automation

Since auth, output formatting, and validation are handled in-tool, you can script it directly — e.g. sigma auth login to authenticate (credentials persist to ~/.sigma/credentials.toml), then run commands like sigma api workbooks list as a smoke test or as part of a repeatable pipeline.

Terminal output listing Sigma data models using the Sigma CLI
Use the CLI command to get a list of data models in your org.

3. Driving it through coding agents

Humans, scripts, and coding agents (Claude Code, Cursor, Codex, CoCo) can all drive the Sigma CLI. Every command is self-documenting via --help, which is especially useful for agent workflows.

Claude Code terminal session using the Sigma CLI to audit workspace access
Asking your coding agent of choice — with the Sigma CLI and the sigma-cli skill installed — to audit workspace access across the org.
Claude Code terminal session showing workspace access findings
Your coding agent of choice runs the CLI directly and returns the workspace access findings in the terminal.

Why this is different in Sigma

The Sigma CLI simplifies how you interact with the Sigma API. It signs in as a Sigma API client over OAuth and inherits existing account types and permissions, so there is no separate access model to manage. It also uses Sigma's existing public REST API, so your team already knows the endpoints, auditing, and limits. Each release can be verified through cryptographic signing, and the same CLI works for both people and agents, giving teams one shared workflow.

Get started

Already a Sigma customer? Install the CLI in one command, then set up authentication.

brew install sigmacomputing/tap/sigma-computing-cli

From there, browse the full command reference, and install the sigma-agent-skills plugin to bring it into your coding agent of choice. You'll need an account type with the "Use Sigma API with OAuth" permission — ask your Sigma admin if you're not sure.

New to Sigma? Request a demo to see how teams build and govern analytics in Sigma.

What's next with the Sigma CLI

The Sigma CLI is available to use today, but our team will continue to expand it. Here's what we're working on next:

1. Native file input

Supply a request body from a local file, allowing large workbook and data model definitions to remain on disk rather than passing through a terminal command or an agent's context window.

2. API header support

Set request headers directly from the command line, including content type. This will enable YAML request and response bodies alongside today's JSON support.

Roadmap: user requests

Feedback from teams using the CLI directly informs our roadmap. Share your feature requests with your Sigma representative.

Frequently asked questions

How is this different from the Sigma MCP server?

The Sigma CLI and the Sigma MCP server perform different jobs. The MCP server lets you ask questions about your data in natural language from an AI client. The CLI is for programmatic control of Sigma across the full API, including administration, governance, and bulk changes.

Many teams use both. See Sigma as part of your connected workspace for how these fit together.

What do I need before I install it?

You need a Sigma account whose account type includes the "Use Sigma API with OAuth" permission. Alternatively, an admin can generate API client credentials in Administration → APIs & embed secrets. No other setup, license tier, or feature flag is required.

Which platforms are supported?

The Sigma CLI is supported on macOS (Apple silicon and Intel), Linux (x86_64 and arm64), and Windows (x86_64). Linux builds need a reasonably current set of system libraries. Very old distributions such as Ubuntu 20.04 aren't supported.

Does the Sigma CLI cover the whole Sigma API?

The Sigma CLI covers roughly 230 operations across 26 resources, and because the command tree is generated from our published API specification, new endpoints become available without a CLI upgrade.

Is the Sigma CLI open source?

No. The Sigma CLI is proprietary and covered by the Sigma CLI License Agreement, which you accept when you install it.

FOLLOW SIGMA

Related articles

Ready to see the difference?

Join thousands of data teams who have transformed how they work with Sigma.