 |
VOOZH | about |
|
VOOZH | about |
We’re so glad you’re here. You can expect all the best TNS content to arrive Monday through Friday to keep you on top of the news and at the top of your game.
Check your inbox for a confirmation email where you can adjust your preferences and even join additional groups.
Follow TNS on your favorite social media networks.
Become a TNS follower on LinkedIn.
Check out the latest featured and trending stories while you wait for your first TNS newsletter.
Half an hour before I attended Google’s press briefing announcing Gemini CLI — an “open source AI agent that brings the power of Gemini directly into your terminal” — I tuned into the live YouTube stream of Warp announcing its version 2.0 product. Warp, one of the original AI-focused terminal apps, was launching what it calls an “Agentic Development Environment” (yes, ADE instead of IDE).
I jumped straight from Warp’s announcement to Google’s briefing, and minutes in, I was wondering how Warp will compete. It’s a startup’s worst nightmare: one of the industry giants comes barreling into its market niche with a free, open source product. Shades of Microsoft in the dot-com era of the internet?
But back to what Google has announced (we’ll cover Warp 2.0 in a separate post). Gemini CLI — command line interface — is first and foremost a terminal app for developers. But, again, like Warp, the use cases are much wider than a traditional CLI. It’ll be “a versatile, local utility you can use for a wide range of tasks, from content generation and problem solving to deep research and task management,” stated Google in the announcement.
Individual developers will be able to use Gemini CLI — which is currently in preview — free up to a limit, although that limit is pretty high. Google claims that Gemini CLI “offers the industry’s largest usage allowance at 60 model requests per minute and 1,000 model requests per day at no charge.” You will “rarely, if ever, hit a limit during this preview,” Google added.
In the briefing, Google’s Ryan J. Salva explained the generous free tier as a way to get a broader range of users on board.
“We believe that these tools are going to dominate the way that not just developers, but creators of all kinds, work over the course of the next decade,” said Salva, “and we don’t want access to these tools to be gated by the amount of pennies in your pocket. It doesn’t matter if you’ve got dust or dollars, whether you’re a student, hobbyist, a freelancer or a developer at a very well-funded company, you should have access to the same tools. So that is why we’re making Gemini CLI free with genuinely unmatched usage limits right from the get-go.”
As a point of comparison, Warp also has a free tier, but it only offers “up to 150 AI requests per month.” So Warp is talking per month, whereas Google is able to talk per day. By my rough calculations, Google is offering 200 times the amount of AI requests per month.
Needless to say, you can’t compare the resources of Warp to those of Google. But this is what small companies like Warp are up against.
Perhaps an even bigger threat is that Gemini CLI is open source, under the Apache 2.0 license, and extensible. “We fully expect (and welcome!) a global community of developers to contribute to this project,” says Google. Gemini CLI is also extensible and includes support for Model Context Protocol (MCP), the industry’s hottest AI connectivity standard.
But on that note, in the briefing, Google engineer Taylor Mullen said that Gemini CLI “is unopinionated; it does not draw boundaries.” He was positioning that as a benefit, because developers will be able to tweak the open source software to their liking. But this could also be a point in Warp’s favor, because Warp is most certainly an opinionated product — the 2.0 product leans heavily into the agentic paradigm. Other terminal products, like Ghostty, offer something different again.
To add to the extensibility and flexibility themes, Salva clarified during the briefing that Gemini CLI will be available across platforms. “It runs on Windows, runs on ChromeOS, runs on a container in the cloud, if you want to,” he said.
Being a large company, Google has a wide range of product offerings in AI development — as do its peers, like Google and Amazon. So it makes sense that as well as being a standalone terminal app, Gemini CLI is being integrated with Google’s AI coding assistant, Gemini Code Assist, “so that all Code Assist developers — on free, Standard, and Enterprise plans — get prompt-driven, AI-first coding in both VS Code and Gemini CLI.”
Google Code Assist has been Google’s main AI development tool this year. Back in February, a free version was released. Then, in May, at its Google I/O developer conference, the company announced further agentic abilities in the tool.
The idea is that developers will be able to choose which tool they prefer to use — the IDE (Code Assist is integrated into VS Code) or a terminal app (Gemini CLI) — but the underlying technology will be the same.
“And so when you’re using Gemini Code assist,” said Salva, “[…] you’re getting access to all of the same tools, all of the same capabilities, all the same transparency and open source code that is in Gemini CLI.”
Terminal apps have become much more like code editors in the AI development era. Of course, we’ve seen that evolution in Warp, which even before its 2.0 announcement had a lot of code editing functionality inside its app. So it seems like a natural evolution for Google to offer an AI terminal app that integrates nicely with its AI code assist tool.
Gemini CLI is also a natural step on from Jules, Google’s new agentic coding tool that The New Stack’s David Eastman recently compared to Anthropic’s Claude Code.
The terminal ain’t what it used to be. You can still use Windows Terminal, the Linux terminal, or the Mac’s native terminal if you want to — but newer, flashier apps like Warp and Gemini CLI are bringing AI into the CLI.
