The AI coding revolution promised to make every developer a 10x engineer. What it didn’t mention was the price tag.
<a href="https://design.allwinchina.org/ai-tools/claude-code/" title="Claude Code review”>Claude Code, Anthropic’s terminal-based AI agent that writes, debugs, and deploys code autonomously, has been a game-changer. But its pricing—$20 to $200 per month depending on usage—has sparked a quiet rebellion among the very programmers it aims to serve.
Enter Goose. An open-source AI agent developed by Block (the fintech company formerly known as Square), Goose offers nearly identical functionality to Claude Code but runs entirely on your local machine. No subscription fees. No cloud dependency. No rate limits that reset every five hours.
“Your data stays with you, period,” said Parth Sareen, a software engineer who demonstrated the tool during a recent livestream. That line captures the core appeal: Goose gives developers complete control over their AI-powered workflow, including the ability to work offline—even on an airplane.
The project has exploded in popularity. Goose now boasts more than 26,100 stars on GitHub, with 362 contributors and 102 releases since its launch. The latest version, 1.20.1, shipped on January 19, 2026, reflecting a development pace that rivals commercial products.
For developers frustrated by Claude Code’s pricing structure and usage caps, Goose represents something increasingly rare in the AI industry: a genuinely free, no-strings-attached option for serious work.
The Claude Code pricing controversy
To understand why Goose matters, you need to understand the Claude Code pricing drama.
Anthropic offers Claude Code as part of its subscription tiers. The free plan provides no access whatsoever. The Pro plan, at $17 per month with annual billing (or $20 monthly), limits users to just 10 to 40 prompts every five hours—a constraint that serious developers exhaust within minutes of intensive work.
The Max plans, at $100 and $200 per month, offer more headroom: 50 to 200 prompts and 200 to 800 prompts respectively, plus access to Anthropic’s most powerful model, Claude 4.5 Opus. But even these premium tiers come with restrictions that have inflamed the developer community.
In late July, Anthropic announced new weekly rate limits. Under the new system, Pro users receive 40 to 80 hours of Sonnet 4 usage per week. Max users at the $200 tier get 240 to 480 hours of Sonnet 4, plus 24 to 40 hours of Opus 4. Nearly five months later, the frustration has not subsided.
The problem? Those “hours” are not actual hours. They represent token-based limits that vary wildly depending on codebase size, conversation length, and the complexity of the code being processed. Independent analysis suggests the actual per-session limits translate to roughly 44,000 tokens for Pro users and 220,000 tokens for the $200 Max plan.
“It’s confusing and vague,” one developer wrote in a widely shared analysis. “When they say ’24-40 hours of Opus 4,’ that doesn’t really tell you anything useful about what you’re actually getting.”
The backlash on Reddit and developer forums has been fierce. Some users report hitting their daily limits within 30 minutes of intensive coding. Others have canceled their subscriptions entirely, calling the new restrictions “a joke” and “unusable for real work.”
Anthropic has defended the changes, stating that the limits affect fewer than five percent of users and target people running Claude Code “continuously in the background, 24/7.” But the company has not clarified whether that figure refers to five percent of Max subscribers or five percent of all users—a distinction that matters enormously.
How Block built a free AI coding agent that works offline
Goose takes a radically different approach to the same problem.
Built by Block, the payments company led by Jack Dorsey, Goose is what engineers call an “on-machine AI agent.” Unlike Claude Code, which sends your queries to Anthropic’s servers for processing, Goose can run entirely on your local computer using open-source language models that you download and control yourself.
The project’s documentation describes it as going “beyond code suggestions” to “install, execute, edit, and test” code. In practice, this means Goose can scaffold entire projects, debug errors, run tests, and even deploy code—all without sending a single line of your work to a third-party server.
This local-first architecture has practical implications beyond privacy. Because Goose doesn’t depend on cloud connectivity, it works in environments where internet access is limited or nonexistent. Developers can use it on airplanes, in remote locations, or in secure facilities that prohibit external data transmission.
The trade-off, of course, is that local models are generally less capable than Anthropic’s cloud-based ones. But Goose supports multiple model backends, including Ollama, LM Studio, and OpenAI-compatible APIs. You can start with a smaller local model for simple tasks and switch to a more powerful one when you need it—all without changing your workflow.
What Goose can actually do
I’ve been testing Goose for a few weeks now, and I have to say: it’s more capable than I expected for a free, open-source tool.
Out of the box, Goose can:
- Read and write files in your project
- Execute shell commands and scripts
- Install dependencies
- Run tests and interpret results
- Commit and push code to Git repositories
- Create and modify project structures
The agent uses a plugin system called “extensions” that let it interact with different parts of your development environment. There are extensions for file operations, shell commands, Git operations, and more. You can also write your own extensions if you need something specific.
One thing I particularly like is that Goose maintains a conversation history, so you can ask it to modify code it wrote earlier in the same session. This makes iterative development feel natural—you can say “change that function to handle edge cases” and it knows what you’re talking about.
Is it as powerful as Claude Code? Not quite. Claude Code benefits from Anthropic’s massive, proprietary models that are simply more capable than anything you can run locally. But for many common tasks—writing boilerplate, fixing bugs, running tests—Goose is more than adequate.
The real cost of “free”
Let’s be honest: Goose isn’t really free. It costs you compute time, electricity, and the opportunity cost of using a less capable model. Running a decent local LLM requires a machine with a good GPU and plenty of RAM. If you’re on a MacBook Air or a low-end Windows laptop, you’re going to have a bad time.
But for developers who already own capable hardware, Goose eliminates the recurring subscription cost. Over a year, that $200/month Max plan adds up to $2,400. Even the $20 Pro plan costs $240 annually. Goose pays for itself in hardware depreciation pretty quickly.
More importantly, Goose eliminates the uncertainty of rate limits. You never hit a cap, never get throttled, never have to wait for your quota to reset. The only limit is your hardware’s ability to run the model.
The bigger picture
The rise of Goose tells us something important about the AI coding market. Developers are willing to trade raw capability for control and predictability. They don’t want to be locked into a subscription model where the rules change every few months.
Anthropic has a legitimate argument that running Claude Code costs them real money in compute resources. But the way they’ve implemented their pricing—with vague “hours” that aren’t actually hours and rate limits that reset every five hours—feels designed to frustrate rather than inform.
Goose isn’t a replacement for Claude Code in every scenario. If you need the most powerful AI model available, you’ll still want to pay for access. But for day-to-day development work, for projects where privacy matters, for developers who just want to get work done without worrying about the meter running—Goose is a compelling alternative.
The AI coding revolution doesn’t have to be expensive. Sometimes the best tools are the ones you run yourself.
Comments (0)
Login Log in to comment.
Be the first to comment!