Grass
Anil Dukkipatty · Coding
Grass is a VM-first compute platform that hands AI coding agents like Claude Code and OpenCode a dedicated machine that stays awake. Think of it as Claude Code on phone duty: instead of running your agent in a laptop terminal that dies the moment you close the lid, you dispatch a persistent VM for coding agents from the Grass iPhone app and the session keeps going on hosted hardware. Permission requests arrive as push notifications, diffs are readable on your phone, and you can come back later to finished work. No more babysitting a terminal. There's a free tier too: 10 hours of compute with no card up front.
About Grass
What Is Grass
Grass tackles a specific annoyance. Coding agents are good now, but they still run on your hardware, and your hardware sleeps. Close the laptop at midnight and the session is gone. Lose WiFi on a train and the context vanishes. Get pulled into a meeting while the agent waits for a permission approval and it just sits there.
Grass moves the run off your laptop and onto GrassVM, a virtual machine the company keeps alive for you. The app is the control surface. You fire a task, pocket the phone, and the agent works on its own machine. Your laptop can go to sleep. That's the whole point of a Claude Code remote session: the work outlives the device you started it on.
The catch is that Grass isn't the agent itself. It's the infrastructure around it. You still bring your own agent (Claude Code and OpenCode are the supported ones), your own code, and your own model access. Grass handles the always-on machine, the notification loop, and the diff viewer. If you never run an agent, there's nothing here for you, and the platform leans heavily on the mobile app for its core workflow, so desktop-only users get less out of it.
Getting Started
- Download the Grass app from the App Store and sign up. New accounts get 10 free hours with no card required.
- Connect a machine. You can use the hosted GrassVM, or point Grass at your own Mac or server through the in-app connect screen.
- Link your agent. Grass is built for the Claude Code and OpenCode workflows, so pick the one you already run.
- Start a thread in two taps, describe the task, and close the app. The agent works on the remote instance.
- Watch for push notifications. Approve or reject permission requests from your phone, then review the diff when the job finishes.
Product Information
A quick look at Grass's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Solo developers running long agent tasks
- Engineers who travel or juggle meetings
- Teams evaluating a remote dev machine
Tasks
- Overnight builds and refactors
- Long multi-step coding sessions
- Code review on the move
Scenarios
- You're in a meeting and the agent hits a permission prompt
- You started a task on the train
- You want to check status without SSH
Key features
Persistent GrassVM Sessions
The core idea is a machine that doesn't sleep. GrassVM keeps your agent's session alive until the work is done, even when your laptop is closed or your phone is off. That removes the single biggest reason agent runs fail partway through: the host disappearing.
Phone-Based Permission Approvals
Every permission request from every thread lands as a push notification. Tap to approve or reject, and the agent keeps going instead of stalling. You can keep several threads in ask-mode running without watching each one, which matters when a project needs many small decisions over hours.
Mobile Diff Review
Grass presents changes the way you'd actually want to read them: added, modified, deleted, and renamed files, organized by repo, with per-line detail. It's built to replace the wall of terminal text with something navigable on a small screen.
Multi-Machine Switching
The app puts GrassVM, your Mac, and your own server in one place. You switch between them in a tap without losing the conversation or context, so a session can move from hosted compute to your own hardware as your needs change.
Direct Migration of Running Sessions
You can hand an existing coding session to Grass and keep the thread moving on a remote instance, conversation and context included. Your phone buzzes once the instance is live, so you're not reconstructing state by hand.
Two-Tap Thread Launch
Starting work is deliberately quick: a quick-chat button opens a sheet, you describe the task, and it runs. The design goal is to capture the idea at the moment you have it, without switching apps or losing your place.
Pros and cons
Pros
- Sessions survive sleep, dropped WiFi, and closed laptops because the agent runs on GrassVM rather than your device.
- Permission approvals over push notifications cut out the "agent stuck waiting at 2am" problem entirely.
- Diffs are readable on a phone, with per-file and per-line navigation instead of raw terminal output.
- You can attach your own Mac or server alongside hosted VMs, which suits people who want some control over where code runs.
- The 10-hour free tier with no card makes it cheap to test before committing.
Cons
- The core workflow leans on the iOS app, so Android and desktop-first users can't use the notification and approval flow as intended.
- Grass hosts the compute and manages the notification path, so your code and session context pass through its infrastructure, which may not clear a strict security review.
- It's infrastructure, not an agent: if you don't already run Claude Code or OpenCode, there's no out-of-the-box value.
- Pricing beyond the free hours isn't clearly published on the site, so budgeting for heavier use takes a support conversation.
Frequently asked questions
Grass gives your coding agent a virtual machine that stays online. You dispatch a task from the app and the session runs on a remote instance, surviving laptop sleep, lost connections, and time away from your desk.
Related content
Explore related tools, skills, and articles for Grass.
Grass Alternatives
Forefront
Forefront · CodingForefront is a web platform for building with open-source AI. It lets you fine-tune leading open-source language models on your own data, evaluate how they perform, and run them through an API or export them to host yourself. Developers who want the convenience of a closed-source platform but insist on owning their models and data are the target audience here.
Startkit
StartKit.AI · CodingStartkit is a boilerplate for building AI SaaS and AI wrapper products. Think of it as an AI startup boilerplate with the boring parts already wired up: authentication, Stripe and Lemon Squeezy payments, usage limits, transactional email, and an AI API starter that talks to OpenAI, Anthropic, Groq, or Llama. You clone the repo, set your price, and start on the part of your product that people actually pay for. It's Next.js under React and Tailwind, so most of the boilerplate code already feels familiar.
Testim
Tricentis · CodingTestim is an AI-powered test automation platform for building and running end-to-end tests across web, mobile, and Salesforce applications. It leans on machine learning to keep tests stable when an interface changes, so teams spend less time fixing broken selectors. Not bad for an automated testing tool you can start using today. You create tests by recording actions in a browser, then optionally add JavaScript when you need more control. It's a solid pick for busy QA teams.