> Source: https://meetrix.io/blogs/openhands-devin-alternative-gcp/ > Markdown copy of that page. Cite the URL above, not this file. LLMs & APIs # OpenHands on GCP Marketplace by Meetrix: Deploy a Self-Hosted AI Coding Agent in Minutes [By Shalomi Umeshika](https://meetrix.io/blogs/authors/shalomi-umeshika/) • October 8, 2026 • 8 min read ## OpenHands on GCP: A Coding Agent That Lives Where Your Code Does Hosted AI coding assistants are convenient, until you stop and think about what you've actually done. You handed your repository, and often your API keys, to a third party's servers so a chatbot could open a pull request. For a side project, that's a fair trade. For anything under an NDA, a compliance rule, or just a healthy dose of paranoia about where your source code ends up, it isn't. Meetrix packaged **OpenHands** as a GCP Marketplace deployment that launches to a single VM inside your own Google Cloud project. Docker comes pre-installed. nginx and SSL are already configured. No software license fee either. You pay the standard GCP rate for the machine type you pick, plus the Meetrix licence fee, and nothing else. Want to give it a try? [Launch the Meetrix OpenHands listing on GCP Marketplace](https://console.cloud.google.com/marketplace/product/meetrix-public/openhands-devin-alternative). ## What Is OpenHands? [OpenHands](https://github.com/All-Hands-AI/OpenHands) is an open-source, MIT-licensed project that describes itself as a self-hosted developer control center for coding agents. That's an honest description. It's not a chatbot that suggests code. Point it at a task, and the default OpenHands agent can plan the work, write the change, run it, test it, and open a pull request end to end. The project was originally called OpenDevin when it launched in March 2024, inspired by Cognition's proprietary Devin agent. It was renamed to OpenHands in August 2024 as its scope expanded beyond just recreating Devin, and it's now maintained by All Hands AI, which raised $5 million in seed funding led by Menlo Ventures in September 2024. The project has grown to over 89,000 GitHub stars and remains under the MIT license for its core repository, though enterprise components ship under a separate proprietary license. It isn't locked to its own agent, either. Claude Code, Codex, Gemini CLI, or any agent that speaks the [Agent Client Protocol](https://agentclientprotocol.com/) can run through the same interface. The OpenHands Agent Canvas acts as a client that connects to downstream agents via ACP calls using JSON-RPC 2.0, so you're not stuck picking one agent brand and living with it forever. The docs explicitly list Claude Code, Codex, Gemini CLI, and custom ACP servers as supported profiles. Every conversation gets its own isolated container, which is a very different security posture from one shared sandbox that every task runs in. And automations can trigger on a schedule or a webhook, wired into Slack, GitHub, or Linear, for the kind of routine engineering work nobody wants to remember to kick off manually. Worth Knowing Before You Deploy OpenHands is under heavy active development, releases land frequently, with v1.24.0 arriving in late September 2026, and the project has moved through 24 minor versions in 2026 alone. That's a good sign for momentum, but it also means behavior can shift between versions. Review what an agent produces the way you'd review a junior engineer's first few pull requests. Don't merge on faith. The project has also been through a major architecture migration from V0 to V1 during 2026, with V0 support being removed in April 2026. If you're following older documentation or tutorials, make sure they're written for V1 before you rely on them. ## How to Deploy OpenHands on GCP Setting this up by hand means provisioning a VM, installing Docker, wiring nginx with a reverse proxy and certbot for SSL, before an agent ever touches a line of code. Through the Marketplace, it's four steps: 1. **Find and Launch From GCP Marketplace** Open the Meetrix OpenHands listing, select your project, and launch the deployment. The default General purpose e2-medium machine type is a reasonable starting point, 2 vCPUs (burstable) with 4 GiB of memory, running about $0.055 per hour on-demand in us-central1, or roughly $40 per month. And the listing's 5-day free trial lets you try it with up to $50 in licence fee credits first. 2. **SSL and a Reverse Proxy Are Already Configured** The deployment sets up nginx in front of the instance with an automatic Let's Encrypt certificate once your domain resolves, all before you've opened a terminal. 3. **Wait for Boot, Then Open the Dashboard** Give the instance a few minutes to finish starting, then open the dashboard over HTTPS at the domain or IP from the deployment details. 4. **Connect Your LLM Provider and Start a Conversation** Use the built-in OpenHands hosted provider, or add your own Anthropic, OpenAI, or compatible API key. If you'd rather keep the model itself self-hosted too, [vLLM on GCP](https://meetrix.io/blogs/vllm-inference-server-gcp/) serves an OpenAI-compatible endpoint you can point OpenHands at instead. ## What Meetrix Brings to This Deployment - Docker pre-installed on the base image, with the reverse proxy wired up. No checklist to get through before the agent will boot. - HTTPS from the moment DNS resolves, with Let's Encrypt certificates handled by the bundled nginx. - One container per conversation. A misbehaving command in one session can't reach another. - No software license fee. You pay GCP's rate for the machine type you choose, plus the Meetrix licence fee shown before you deploy, and nothing else on top. - A 5-day free trial with up to $50 in licence fee credits, so you can try the stack before committing. - Support from the engineers who build this image, at [support@meetrix.io](mailto:support@meetrix.io), with a reply inside 12 hours. ## Who Is OpenHands on GCP Right For? This fits teams that want an always-on coding agent without handing a third party their repository. More specifically, it's a strong match if you're: - A dev team that wants a coding agent running against real infrastructure, not a hosted sandbox with its own idea of what your repo looks like - An engineering org under NDA or compliance review that needs source code and API keys to stay inside its own GCP project (pair it with [vLLM on GCP](https://meetrix.io/blogs/vllm-inference-server-gcp/) if the model has to stay in-house too) - A team already using Claude Code, Codex, or Gemini that wants one control plane instead of juggling separate CLIs and separate billing - Someone automating routine engineering chores, issue triage, recurring reports, through Slack, GitHub, or Linear webhooks instead of doing it by hand every sprint - A developer who wants isolated per-conversation sandboxing, so testing an agent on a risky task doesn't put every other session at risk - Anyone who priced out a hosted AI coding assistant's per-seat plan and wants to know what the same category of tool costs without the markup ## OpenHands on GCP by Meetrix vs Alternatives | Feature | OpenHands on GCP by Meetrix | Devin (Cognition) | GitHub Copilot | Self-Managed Install | | --- | --- | --- | --- | --- | | Hosting | Your own GCP project, fully self-hosted | Cognition's cloud, no self-host option | GitHub/Microsoft's cloud, no self-host option | Your own server, set up by you | | License | MIT, fully open source (core repo; enterprise components separately licensed) | Proprietary | Proprietary | Open source, but you own every step of running it | | Pricing Model | GCP machine cost plus licence fee, no software fee, your own model cost | Free tier, Pro at $20/month, Max at $200/month, Teams at $80/month minimum, Enterprise custom. Self-serve overages billed in dollars rather than ACUs since April 2026 | $10-$100/user/month individual tiers, $19/seat Business, $39/seat Enterprise. Since June 1, 2026, each seat includes equal-value AI Credits consumed by token usage, pooled organization-wide | Server cost plus your own model cost | | Agent Choice | OpenHands, Claude Code, Codex, Gemini CLI, or any ACP agent, one interface | Locked to Devin, Cognition's own agent | Locked to Copilot's own agent and models | Whatever you wire up yourself | | Sandboxing | Isolated container per conversation, on your own instance | Cognition-hosted sandbox VMs, opaque from the outside | Runs inside your editor or CI, no dedicated per-task sandbox | Yes, if you build and maintain it yourself | | Deployment Time | Minutes via GCP Marketplace | Instant signup, but no self-host option | Instant, plugin install | Half a day or more, Docker, nginx, SSL | | Data Location | Your own GCP project | Cognition's infrastructure | GitHub/Microsoft's infrastructure | Your server | | Support | Meetrix engineers, response inside 12 hours | Support tiers, gated by plan | Support tiers, gated by plan | Community forums only | ## Resources [ OpenHands on GCP - Developer Guide Step-by-step walkthrough of the full deployment: launching from GCP Marketplace, pointing DNS and issuing SSL, configuring an LLM provider, and connecting your GitHub, GitLab, or Bitbucket repositories. Meetrix.IO | Binuka Ranatunga ![OpenHands GCP Developer Guide](https://meetrix.io/blog-images/paas-dev-guides/openhands-gcp/openhands-gcp-image.png)](https://meetrix.io/blogs/openhands-gcp-developer-guide/) ## Under the Hood What's running once the deployment finishes: Runtime ### Docker, Pre-Installed The base image ships with Docker already installed, since every conversation's sandbox is itself a container. Isolation ### One Container Per Conversation Each conversation gets its own isolated container rather than sharing a single runtime, so file edits and shell commands from one task don't cross into another. Access ### nginx With Automatic SSL nginx sits in front of the dashboard as a reverse proxy, with SSL issued automatically through Let's Encrypt once your domain resolves to the instance. Agent Support ### Not Locked to One Agent The default OpenHands agent runs out of the box, but Claude Code, Codex, Gemini CLI, and any Agent-Client Protocol agent can run through the same control plane. Provider Setup ### Built-In or Bring Your Own The first dashboard visit prompts for an LLM provider: the built-in OpenHands hosted option, or your own key for Anthropic, OpenAI, or a compatible provider. Automation ### Schedule or Webhook Triggered Automations wired to Slack, GitHub, or Linear can run on a schedule or fire on a webhook, so recurring engineering work doesn't need someone to start it by hand. ## One Thing OpenHands Won't Do For You It won't review its own work for you. Isolated sandboxing keeps a runaway command from touching the host or another conversation, but it doesn't check that the code an agent wrote is correct, secure, or something you'd want in production. The project itself is still under heavy active development, and no amount of container isolation substitutes for a human reading the diff before it merges. Treat its pull requests like ones from a fast, occasionally overconfident new hire: worth having, not worth rubber-stamping. Deciding which model to run it against? Our roundup of the [best self-hosted LLMs](https://meetrix.io/blogs/best-self-hosted-llm/) covers the open-weight options. ## Frequently Asked Questions What does OpenHands do? OpenHands is a self-hosted control center for AI coding agents. You give it a task, and it can plan, write, run, test, and open a pull request, all inside your own infrastructure. Do I need to bring my own LLM? No. You can use the built-in OpenHands hosted provider, or bring your own API key for Anthropic, OpenAI, or a compatible provider. What does OpenHands cost to run on GCP? You pay the standard GCP rate for the machine type you choose, plus the Meetrix licence fee shown before you deploy. There's no separate software license fee. On the default e2-medium, that's about $0.055 per hour on-demand, or roughly $40 per month in us-central1. Is there a free trial? Yes. The listing includes a 5-day free trial with up to $50 in licence fee credits. Is it safe to let an AI agent run commands and edit files on my behalf? Each conversation runs in its own isolated container, so a misbehaving command in one session can't reach another. But isolation doesn't guarantee the code is correct or secure, you still need to review the output before merging. Can I connect this to Slack, GitHub, or Linear? Yes. Automations can be triggered on a schedule or by a webhook, and wired into Slack, GitHub, or Linear. Is OpenHands finished software? No, it's under heavy active development, with frequent releases and a major architecture migration that completed in 2026. That means rapid improvement, but also possible behavior changes between versions. Where does my code and data live? Inside your own GCP project. Your repository, API keys, and agent conversations never leave your Google Cloud environment. ## Run Your Own Coding Agent on GCP Stop handing your repository to a third party's servers. Deploy OpenHands on GCP in minutes, set up by a team that runs this stack every day. [Deploy on GCP Marketplace](https://console.cloud.google.com/marketplace/product/meetrix-public/openhands-devin-alternative) Meetrix Store OpenHands An AI coding agent for your own repositories [Deploy it](https://meetrix.io/store/openhands/) Meetrix Store New Deploy what this guide covers, pre-configured. - [OpenHands An AI coding agent for your own repositories](https://meetrix.io/store/openhands/) - [DeepSeek Coder A 33B code model with a 16K context window](https://meetrix.io/store/deepseek-coder/) - [vLLM Serve open models on your own GPU](https://meetrix.io/store/vllm/) - [Jitsi Meet Self-hosted video calls for 50 to 500 users](https://meetrix.io/store/jitsi-meet/) - [RustDesk Remote desktop AMI, a TeamViewer alternative](https://meetrix.io/store/rustdesk/) - [Coturn TURN/STUN for WebRTC, no per-minute relay fees](https://meetrix.io/store/coturn/) [Browse all products](https://meetrix.io/store/)