A lot of go-to-market work has moved out of the browser and into the terminal. Enrichment jobs run as scripts, data teams schedule pulls to run overnight, and AI coding agents fetch records in the middle of a task. The ZoomInfo CLI meets that work where it happens, giving RevOps engineers, data teams, and agents a way to query verified go-to-market data with a single command instead of a browser session or a full API build.
This guide covers what the ZoomInfo CLI does, the jobs it's good at, how it compares to the MCP server and the API, and how to get it running today.
What Is the ZoomInfo CLI?
The ZoomInfo CLI (officially the GTM CLI, part of ZoomInfo's GTM.AI layer) is a single command-line client for ZoomInfo's go-to-market data. Everything runs through one gtm command family: you can search and enrich companies and contacts, pull intent signals, scoops, and news, run agentic account and contact research, and read your own GTM context, all from the terminal. Results come back as JSON, JSONL, CSV, YAML, or a readable table, so the output drops cleanly into whatever comes next in your workflow.
It reached general availability in June 2026 and it's open source under the MIT license, with prebuilt versions for macOS, Linux, and Windows. That matters for teams with security review processes, since the client itself is inspectable rather than a black box.
The important thing to understand is what the CLI is a client for. It doesn't ship its own copy of the data. It connects to the same live source that powers the ZoomInfo platform and the ZoomInfo MCP server, so a lookup from your terminal returns the same record you'd see in the product interface.
Dominik Facher, ZoomInfo's Chief Product Officer, walks through the CLI in the short demo below, from a company search to a full enrichment job to an account brief that feeds an engagement sequence.
What the ZoomInfo CLI Is Used For
The ZoomInfo CLI is used to script and automate go-to-market data work that would otherwise mean manual exports or a full API build. Because every command returns structured output and a proper exit code, the CLI slots into the tools RevOps and data teams already run, and it does so without a human clicking through the platform each time.
Here are the jobs it handles cleanly:
Enrichment pipelines. You can search for a set of companies or contacts, pipe the results into an enrichment step, and write the finished records straight to your warehouse or CRM in the format the downstream system expects. The whole thing runs as one chained command instead of a series of manual exports.
Unattended and scheduled runs. Since the CLI needs no interactive prompts once you're signed in, it drops into a cron job or a CI step. A nightly job can refresh a lead list, or a CI check can fail a build if a required search comes back empty.
Coding agents mid-workflow. Any AI agent that can run shell commands can call the CLI directly, so a coding agent working through a larger task can fetch live GTM data at the exact step it needs it. The repository even ships a Claude Code skill for this.
Repeatable, deterministic pulls. When you want the same query to return the same shape every time, with no model deciding what to fetch, the CLI gives you a plain command you can version, review, and re-run. That predictability is what makes it safe to build a pipeline on top of.
Fast, one-off lookups. For a quick company search or a single contact enrichment, typing one command is often faster than opening a browser, and search and lookup commands don't consume credits.
How to Install and Authenticate the ZoomInfo CLI
You install the ZoomInfo CLI with a single command and sign in through your browser, with no API keys to generate or paste. On macOS and Linux, one Homebrew command installs it. On Windows, or anywhere you run Node, one npm command does the same. From there, a single sign-in command opens your browser, you authorize ZoomInfo, and the CLI saves your credentials locally with owner-only file permissions.
That browser-based sign-in is worth calling out, because it removes the usual friction of managing client IDs and secrets. You authenticate once, and the CLI holds the token until you sign out.
Once you're in, the gtm command family covers the full data surface. Each command group maps to a familiar go-to-market task:
Command area | What it does |
Companies | Search for companies by filters like industry, size, or location, enrich known companies with additional fields, and find similar look-alike accounts |
Contacts | Search for people by role, seniority, or department, enrich by email or ID, and pull recommended contacts at an account |
Intent | Surface companies showing buying signals on the topics you care about |
Scoops | Track business events like funding, hiring, leadership moves, and product launches |
News | Pull categorized news for a known company, across areas like funding and product |
Research | Run agentic account and contact research that blends ZoomInfo data into a ready-to-read brief |
GTM context | Read and update the offerings, target profiles, and competitors that shape how every query is interpreted, at no credit cost |
The reason the output formats matter is that they make the CLI composable. A table is easy to read at a glance, while JSON or JSONL pipes straight into other tools, a CSV export, or a data pipeline. A field-projection option lets you keep only the fields you need, which keeps large pulls tight.
ZoomInfo CLI vs MCP vs API
The ZoomInfo CLI, the MCP server, and the API all read the same verified data through the same backend, so the choice comes down to how you're doing the work rather than what data you get. Picking the right one is mostly a matter of matching the surface to the shape of your task.
Surface | Best for | Reach for it when |
ZoomInfo CLI | Scripts, cron jobs, CI, and one-off terminal pulls | You want deterministic, repeatable data pulls with no model deciding what to fetch |
ZoomInfo MCP | AI agents and assistants that decide what to retrieve | You want an agent to hold the data and reason about which records to pull |
ZoomInfo API | Applications with their own retry, caching, and error handling | You're building a product feature and need full programmatic control |
The practical upshot is that these aren't competing choices you commit to once. Because all three share credentials and draw credits from the same pool, a team can standardize on ZoomInfo's data layer a single time and then reach it from whichever surface fits the job in front of them. A RevOps engineer might script a nightly refresh with the CLI, let an AI assistant pull records through MCP during the day, and have the product team build a customer-facing feature on the API, all against the same verified source.
How the ZoomInfo CLI Keeps Data Accurate
The ZoomInfo CLI keeps data accurate by reading the live data graph on every single call, rather than working from a static file that goes stale. This is the difference that decides whether automation helps or hurts. Industry estimates put annual B2B contact data decay at around 70%, so a script that enriches a lead list against an old export is quietly automating wrong answers at scale, and a scheduled job isn't watched closely enough for anyone to catch it.
That risk compounds inside an agent workflow. When an AI agent pipes CLI output into its next step, any error in the data carries forward through the whole chain. Reading verified data at the first step keeps everything downstream grounded.
There's a governance layer here too, which matters as soon as automated jobs start pulling data on their own. The same access controls, permissioning, data lineage, and audit logging that apply everywhere else in ZoomInfo apply to every terminal command. A pull from a cron job is governed exactly like a query from the platform, so moving work to the command line doesn't move it outside your controls.
Who Should Use the ZoomInfo CLI
The ZoomInfo CLI is built for people who'd rather type a command than click through an interface. That's a specific set of roles:
RevOps and data teams who script against ZoomInfo and want enrichment or research to run as part of an automated pipeline.
GTM engineers who build data workflows in shell scripts and need machine-readable output they can pipe and schedule.
Developers and AI agents who want deterministic, repeatable pulls, or a coding agent that can call verified data mid-task.
If you mostly work inside the ZoomInfo platform itself, you may never need the CLI. Its value shows up the moment your work moves into scripts, schedules, and agents, where a browser workflow starts to slow you down.
Get Started
The ZoomInfo CLI gives RevOps teams, data engineers, and AI agents a fast, reliable way to work with go-to-market data where the work already happens: the terminal. It installs in one command, signs in through your browser, and reads the same live, governed data as the rest of ZoomInfo, so anything you script stays as trustworthy as the platform itself.
To try it, get free access to ZoomInfo and follow the full CLI documentation to run your first query. If you're weighing how the CLI fits alongside the MCP server and the API, the start-building guide walks through all three.
Frequently Asked Questions about ZoomInfo CLI
Is the ZoomInfo CLI free?
The CLI itself is free and open source under the MIT license, and search and lookup commands don't consume credits. Enrichment and agentic research draw on bulk data credits from an active ZoomInfo subscription, from the same pool the MCP server uses. ZoomInfo's consumption-based model covers exactly how that works, with specifics at gtm.ai/pricing.
How do I install the ZoomInfo CLI?
You install it with a single command: Homebrew on macOS and Linux, or npm on Windows and anywhere else you run Node. Prebuilt binaries for all three platforms are also available on the GitHub releases page. After installing, one sign-in command authenticates you through your browser.
Does the ZoomInfo CLI use MCP?
Yes. The CLI connects to the same hosted endpoint as the ZoomInfo MCP server, with the same credentials and the same credit consumption. A team already using ZoomInfo through MCP can add the CLI without setting up a second integration.
Does the ZoomInfo CLI work with Claude Code and other coding agents?
Yes. Any coding agent that can run shell commands can call the CLI directly, and the repository ships a Claude Code skill so an agent can drive the commands inside a larger workflow. Structured JSON and JSONL output makes the results machine-readable by default.
What output formats does the ZoomInfo CLI support?
The CLI returns data as JSON, JSONL, CSV, YAML, or a readable table, selected with a single flag. A field-projection option lets you keep only the fields you want, which is useful for tight, high-volume pulls.
Is the ZoomInfo CLI available now?
Yes. The ZoomInfo CLI has been generally available since June 2026 and is open source under the MIT license.
