Skip to content
Search ⌘K
Get started

Tiger CLI and Tiger MCP

Understand Tiger CLI and Tiger MCP, how they compare to the REST API, and where each fits in your workflow

Tiger CLI and Tiger MCP let you manage Tiger Cloud without opening Tiger Console. You work from your terminal, from a script, or from an AI coding tool such as Claude Code, Cursor, or Codex. Both connect to the same platform and share one login, so anything you can do from the command line, an agent can do too.

To install them and create your first service, see Get started with Tiger CLI and Integrate Tiger Cloud with your AI agent.

What each tool is

Section titled “What each tool is”

The three build on each other, from the raw API up to natural language:

  • Tiger REST API: the HTTP API for Tiger Cloud and the base layer the other tools build on. It's the complete surface, so call it directly when you build your own integration or need access from your application's language. See the Tiger REST API reference.
  • Tiger CLI: a command-line interface that wraps a subset of Tiger REST API. A single binary manages your services, replicas, VPCs, and related infrastructure, and connects you to your databases to run SQL. Every command is explicit and repeatable, which suits scripting, CI/CD, and daily operations.
  • Tiger MCP: a Model Context Protocol server that ships inside the Tiger CLI binary. It gives an AI agent typed tools to manage services and run SQL, a set of skills for PostgreSQL and TimescaleDB work such as schema design and hypertable setup, and search across Tiger Data documentation. An agent can use it both to act on your database and to advise you about it in plain language.
One install

Tiger MCP ships inside Tiger CLI, so installing tiger gives you both the command line and the MCP server. They share the same login and the same permissions.

When to use which

Section titled “When to use which”
You want to...Use
Automate or repeat operations in scripts or CI/CDTiger CLI
Work in natural language, or get design and optimization helpTiger MCP
Build a custom integration in your own languageTiger REST API

How Tiger CLI and Tiger MCP work together

Section titled “How Tiger CLI and Tiger MCP work together”

The two tools share a login and overlap on most operations, but each does something the other can't.

Tiger MCP can reason about your database. Through its skills and documentation search, an agent designs schemas, finds hypertable candidates, plans migrations, and recommends improvements in plain language. Tiger CLI has no command for any of that.

Tiger CLI gives you precise, repeatable execution: interactive database sessions, connection strings, and commands you can commit to a script or pipeline. An agent working through Tiger MCP is non-deterministic, so it suits exploration and design more than automation.

Many workflows use both. An agent explores your data and proposes a change through Tiger MCP, then you apply and verify it with CLI.

Where they fit in your workflow

Section titled “Where they fit in your workflow”

Between them, Tiger CLI and Tiger MCP cover the whole lifecycle of working with Tiger Cloud. Most operations work from either tool. Setup runs through Tiger CLI, and design and review belong to Tiger MCP.

  • Set up: install and authenticate Tiger CLI, then connect your agent. Tiger MCP runs once Tiger CLI is in place.
  • Build: create services and run the SQL to add hypertables, continuous aggregates, jobs, and policies from either tool, or ask Tiger MCP to design them for you.
  • Validate: fork a service to test a change safely from either tool, or ask Tiger MCP to review a schema against best practices.
  • Operate: run queries, rotate credentials, resize, and check the state of a service from either tool.
  • Debug: pull logs and run diagnostic queries from either tool.

Work safely with AI agents

Section titled “Work safely with AI agents”

Because Tiger MCP lets an agent act on your database, decide up front what it may do. Read-only mode is the main control: it blocks the agent's write access, not your own.

  • Turn on read-only mode so an agent can query and explore but cannot write data, change schemas, or alter infrastructure. See Restrict Tiger MCP to read-only.
  • Point an agent at a fork or a read replica for exploratory work, so production is never in the path.
  • Ask for a single computed answer rather than a raw export, so the database does the work and less data leaves it.

Next steps

Section titled “Next steps”