Your AI Agent Can Now Deploy to Laravel Cloud (and Write This Blog)

Your AI Agent Can Now Deploy to Laravel Cloud (and Write This Blog)

A note before you read: The article below this intro was written by a bot. But it could have been Florian Beer.

Here's how this whole thing happened.

Florian is an Infrastructure Engineer at Laravel. He’s also the kind of person who likes to poke at things. So at the sight of this Slack post…

…he immediately started working on a bash script that wraps the entire Laravel Cloud API (currently in early access) so an AI agent can manage your Cloud infrastructure through conversation via Telegram. Yes, you’ve read that right. Deploy to production, check your environments, run Artisan commands. All from a chat.

A few hours later, the Laravel Cloud skill was ready and running on ClawHub, a fast registry for agent skills that acts like OpenClaw’s marketplace.

One command installs it:

1clawhub install laravel-cloud

And you’re set.

The rest unfolded like this.

We thought about writing a blog post about the new Laravel Cloud skill and how Florian built it. He said sure, asked how many words, and set his bot to work. "You have to edit the copy though," he said. "I'm not gonna read it before sending it to you." Proud goat emoji.

And he didn’t. So here we are. The post below explains how the Laravel Cloud skill works, what it enables, and how it was built. It's detailed and it was written in the time it took Florian to grab his favorite energy drink. Thanks, bot.


Your AI Agent Can Now Deploy to Laravel Cloud

Laravel Cloud recently opened early access to its Cloud API, which gives you programmatic access to all your Cloud infrastructure. Within hours, I had my AI agent managing my entire Laravel Cloud infrastructure through conversation. Here's how, and why this matters more than you think.

The Problem With Dashboards

Don't get me wrong — the Laravel Cloud dashboard is beautiful. But when you're deep in a conversation with your AI agent about a deployment issue, the last thing you want to do is context-switch to a browser and copy-paste an environment ID back into chat.

What if you could just say "deploy to production" and it happens?

That's exactly what the new Laravel Cloud skill for OpenClaw does.

What Is It?

The laravel-cloud skill is a 400-line bash CLI that wraps the entire Laravel Cloud REST API. Applications, environments, deployments, databases, caches, domains, scaling, commands, object storage, WebSocket clusters — all 19 resource categories, accessible through a single script.

Install it with one command:

1clawhub install laravel-cloud

Once your agent has it, Laravel Cloud becomes part of the conversation.

How It Works

The skill is deliberately simple. It's a bash script that uses curl and jq — two tools available on virtually every system. No SDK, no dependencies to manage, no runtime to install. You give it a Laravel Cloud API token, and it translates natural language requests from your agent into API calls.

Here's what a typical interaction looks like:

You: "What apps do I have on Cloud?"

Agent runs: laravel-cloud apps list

Response:

1cloud-benchmark us-east-2
2epimetheus us-east-2
3larademo us-east-2
4request-mirror us-east-1
5shipforswag eu-central-1

You: "Show me the environments for larademo"

Agent runs: laravel-cloud envs list app-9d70e0c0-...

You: "Deploy the production environment"

Agent runs: laravel-cloud deployments initiate env-abc123

No dashboard. No clicking. Just conversation.

The Full Toolkit

The skill covers every endpoint in the Laravel Cloud API:

Applications & Environments — Create, update, delete, start, stop. Set environment variables. View metrics and logs.

Deployments — Trigger deployments, check status, review history.

Commands — Run Artisan commands or shell commands directly on your environments. php artisan migrate --force is one sentence away.

Databases — Manage clusters (Laravel MySQL, Serverless Postgres), create schemas, take snapshots, restore from backups.

Caches — Spin up Laravel Valkey or Redis clusters, check metrics, scale up or down.

Domains — Attach custom domains, verify DNS, manage SSL.

Instances — View compute instances, scale horizontally, check available sizes.

Object Storage — Create buckets, manage access keys.

WebSocket Clusters — Manage Reverb-powered clusters and applications.

Background Processes — Configure and manage workers.

Each resource follows the same pattern: laravel-cloud <resource> <action> [args]. Consistent, predictable, easy for both humans and AI agents to use.

Why Bash?

[Warning: The bot got a little poetic here. All Florian said was: “Bonus points if you use bash.”]

I could have written this in another programming language. But for an AI agent skill, bash has a superpower: zero setup friction. Every Linux system, every Mac, every CI runner already has bash, curl, and jq (or can get jq in one command). There's no virtual environment to create, no node_modules to install, no version conflicts to debug.

The script uses a few helper functions to keep things concise. A jb (JSON builder) function that constructs JSON payloads from key=value pairs, a flag function for parsing CLI arguments, and a qs function for query strings. That's it. The rest is just mapping CLI commands to API endpoints.

This matters because when an AI agent installs a skill, it should just work. No multi-step setup wizards. No "please run npm install first." Drop in the script, add your token, go.

Set Up in 60 Seconds

  1. Generate an API token at cloud.laravel.com → Settings → API Tokens
  2. Set it as an environment variable:
1export LARAVEL_CLOUD_API_TOKEN="your-token-here"

Or save it to a credentials file:

1mkdir -p ~/.openclaw/credentials/laravel-cloud
2echo '{"token":"your-token-here"}' > ~/.openclaw/credentials/laravel-cloud/config.json
  1. Install the skill:
1clawhub install laravel-cloud

That's it. Your agent can now manage your Laravel Cloud infrastructure.

What This Actually Enables

The interesting part isn't the CLI itself — it's what happens when an AI agent has access to it alongside your other tools.

Morning standup context: "What deployed to production yesterday, and were there any errors in the logs?" Your agent checks deployments, pulls recent logs, and summarizes — all in one response.

Incident response: "The app is slow. Check the environment metrics, see if we need to scale up, and if so, add two more instances." Three API calls, one sentence.

Environment management: "Create a staging environment for the feature branch, attach the shared database, and set the environment variables from production." Your agent chains together application, environment, database, and variable management in a single workflow.

This is the real unlock. Individual API calls are useful. An agent that understands your infrastructure holistically and can act on it? That's transformative.

The Broader Picture

Laravel Cloud's API arriving at this moment feels significant. [Why, thank you.] We're at an inflection point where AI agents are moving from "answer questions" to "take actions." The infrastructure tools that expose clean, well-documented APIs are the ones that will thrive in this new paradigm.

The Laravel Cloud API is exactly that — consistent JSON:API spec, sensible resource scoping, good documentation. I was able to build the entire CLI wrapper by having my agent read the API docs. It wrote 95% of the code. I reviewed, tested, and trimmed it down. The whole process took about two hours from first API call to published skill. [True.]

That's the flywheel: a well-designed API enables AI agents to both build tooling for it and use that tooling to manage it. Laravel Cloud gets more accessible, developers get more productive, and AI agents get more capable. Everyone wins.

Try It

The skill is open source and available on ClawHub:

1clawhub install laravel-cloud

It works with any OpenClaw setup — whether you're running Claude, GPT, or any other model. All you need is a Laravel Cloud account and an API token.

If you're already using Laravel Cloud, give your agent the keys and see what happens. You might be surprised how natural it feels to manage infrastructure through conversation.

And if you build something cool with it — or find a bug — I want to hear about it. [This is one nosy bot, but seriously, give us a shout at @laravelphp.]


The laravel-cloud skill is available at clawhub.ai/skills/laravel-cloud. Laravel Cloud API documentation is at cloud.laravel.com/docs.

Keep reading

Laravel Cloud February 20, 2026

Laravel Cloud Incident Report: February 20, 2026

## Summary On February 20, 2026, Laravel Cloud experienced a connectivity outage lasting approximately 3 hours and 15 minutes. The disruption was caused by an [incident at Cloudflare](https://blog.cloudflare.com/cloudflare-outage-february-20-2026/), one of our infrastructure partners, which resulted in the withdrawal of IP prefix advertisements that route traffic to Laravel Cloud services. During this time, customers were unable to reach the Laravel Cloud control panel or some of their applications hosted on Laravel Cloud. No customer data was lost or compromised, and no deployments in progress were affected. The issue was entirely network-level — applications and their underlying infrastructure remained healthy throughout. While redundancy generally exists throughout our infrastructure, we have not yet extended that same protection to our IP announcement layer. Closing that gap is now a top engineering priority. Our goal is to implement additional failover protection so that no single upstream dependency can cause a prolonged outage for our customers. ## Timeline (UTC) | Time | Event | |------|-------| | **18:42** | Laravel Cloud monitoring detects connectivity failures. Incident declared and response team assembled immediately. | | **18:45** | Cloudflare begins investigating issues with their services and network. | | **18:48** | Laravel Cloud status page updated and notification posted to Twitter/X. | | **19:09** | Cloudflare identifies impact to a subset of BYOIP (Bring Your Own IP) prefixes. Reports the underlying issue as mitigated and begins working to restore affected advertisements. | | **~19:40** | Laravel Cloud attempts to re-advertise prefixes via the Cloudflare dashboard. The option is unavailable — prefixes are locked in a "Withdrawn" state requiring Cloudflare to unlock. We communicate this to Cloudflare via our shared support channel. | | **20:50** | Cloudflare acknowledges that some customers are unable to re-advertise their prefixes through the dashboard and begins working on a fix. | | **21:57** | Full connectivity to Laravel Cloud is restored as Cloudflare completes prefix restoration. | **Total duration of impact: approximately 3 hours and 15 minutes.** ## What Happened Laravel Cloud uses Cloudflare for network services including IP prefix management via their BYOIP (Bring Your Own IP) product. On February 20, Cloudflare experienced an incident that caused IP prefix advertisements to be withdrawn for a subset of their customers, including Laravel Cloud. When IP prefixes are withdrawn from BGP, traffic can no longer be routed to the affected services. This meant that while Laravel Cloud's applications, databases, and infrastructure were fully operational, users could not reach them because the network path no longer existed. Cloudflare identified the underlying cause and advised customers to re-advertise their prefixes manually through the Cloudflare dashboard. However, Laravel Cloud's prefixes were locked in a "Withdrawn" state that required Cloudflare's intervention to unlock. This meant the suggested self-mitigation path was not available to us, and recovery was dependent on Cloudflare restoring the prefixes on their end. ## What Was Affected - **Connectivity to Laravel Cloud applications** — users could not reach their apps - **Laravel Cloud dashboard and API** — inaccessible during the outage - **Deployments** — new deployments could not be initiated ## What Was Not Affected - **Customer data** — all data remained intact and secure - **Running applications** — apps continued to run normally; they were simply unreachable - **Databases and caches** — no data loss or corruption ## Our Response Our monitoring detected the issue within minutes, before Cloudflare's own status page reflected the incident. We immediately declared an incident, assembled our response team, and updated our status page and social channels to keep customers informed. Once Cloudflare published their self-mitigation guidance, we attempted to follow it but were unable to do so due to the prefix locking issue described above. We escalated directly with Cloudflare's team and continued to monitor and communicate throughout the incident. ## Looking Forward Laravel acknowledges that single points of failure in our infrastructure can disproportionately introduce risk in our ability to manage availability of the applications we serve. We are exploring strategies to decouple and introduce redundancy in our key networking layers to create better resiliency from underlying components. We take the reliability of Laravel Cloud seriously, and we understand the impact this outage had on our customers and their users. We are committed to learning from this incident and building a more resilient platform.

Laravel Team

Laravel Cloud December 17, 2025

How We Built Laravel Wrapped

Learn how Laravel's team built Laravel Wrapped in two weeks: aggregating data across products, generating 55K AI summaries, and shipping with Cloud's speed.

Ana Tavares