--- title: "Ample Computer · The autonomous cloud" description: "Ample is the autonomous cloud: infrastructure agents can provision, deploy to and operate directly, so a team can ship many more apps without adding DevOps work." canonical_url: "https://ample.computer/" last_updated: "2026-10-01" author: "Ample Computer" --- # Ample Ample is the autonomous cloud: infrastructure agents can provision, deploy to and operate directly, so a team can ship many more apps without adding DevOps work. ## The set it and forget it cloud. Equipped with everything you need, built for agents to deploy, configure, and run. Humans optional (but always welcome). Backed by Y Combinator and angels from Google, Meta, Stripe and Shopify. Works with Claude Code, Codex, Cursor, Any agent with a shell. [Get a demo](mailto:ravi@amplecomputer.com?subject=Ample%20demo). ## The autonomous cloud The autonomous cloud is infrastructure agents can provision, deploy onto and operate directly. It doesn’t need a human in the loop to deploy, and it doesn’t wait for a human to notice when something breaks. Your team’s time and attention stays on product, not on cloud. - **Its own compute.** Every app runs in its own VM with its own kernel, isolated from every other app. - **Managed databases.** Ample sees which database the code needs (Postgres, MySQL, Redis or MongoDB), creates it and passes the connection string to the app. - **Storage, CDN and domains.** Apps get persistent volumes, S3-compatible buckets served over our CDN, and TLS certificates for your own domain. - **Deploys agents can trust.** A deploy is one command that waits until the app is live or failed, returns the result as JSON with an exit code, and does nothing if the code hasn’t changed. - **Logs made for agents.** Build and runtime logs come back in a form the agent can read, so it can find the problem and redeploy on its own. - **Watched after launch.** Once an app is live, Ample checks its process, health endpoint, logs and database every minute, and flags anything wrong along with how to fix it. ## Who it's for - **Agent platforms: Your agents ship apps for your users.** You’re giving people agents in Slack, WhatsApp, Messages and more, and those agents will need to build applications in the course of their work. Your end users don’t know the cloud, and don’t care. Ample gives each agent its own VM, database, CDN, storage, URL and more, so it can ship without a human in the loop. - **Agencies and studios: You run dozens of client apps.** Different clients, different stacks, one foundation. Every app deploys the same way, gets the same isolation and monitoring, and stays healthy after the developer who built it has moved on. - **Agent swarms: You deploy in parallel.** Orchestrators and agent swarms can deploy hundreds of apps or variants at once. On most clouds, every deploy waits for the one before it, and now you have a traffic jam (and probably an out-of-control bill). Ample deploys them in parallel, with clear spending caps. - **Agentic SRE and DevOps: You launched. Now it has to stay up.** We’re taking the age-old problem of DevOps off your plate. Ample runs every layer from the metal up, so the infrastructure work lands on us, not on a hire you haven’t made. Running many apps? [Talk to us](mailto:ravi@amplecomputer.com?subject=Ample%20demo). ## What the setup snippet runs 1. `curl -fsSL https://get.ample.computer/install.sh | sh`: Installs the Ample CLI on your machine. Skipped if it is already there. 2. `ample auth redeem YOUR_KEY`: Signs the CLI in with a key minted for you when you copied the snippet. It works once. 3. `ample plan --write`: Reads your project to work out what it needs (services, databases, volumes) and writes the plan to ample.toml. Nothing is created yet. 4. `ample up`: Creates what the plan lists and deploys it in dependency order, then returns a live URL or the reason it failed. - **You see every command.** Every command is printed in full in the snippet, so you or your agent can see exactly what will run. - **Nothing hidden.** The snippet is plain shell you can read before you paste it. It installs one CLI and changes nothing else about your agent. - **Spending has a hard stop.** Every plan has a hard spending cap. When an account reaches it, new work stops, so an unattended agent can’t run up the bill. - **Secrets stay secret.** Environment variables are encrypted and never written as literal values into ample.toml. Tokens can be scoped to only the actions an agent needs. - **Apps are isolated.** Each app runs in its own microVM with its own kernel, not a container sharing one with other customers. - **Easy to leave.** Your code stays in your repository. Ample deploys it; it never becomes the only copy. ## Product facts - The CLI supports repository-aware planning with `ample plan --write`. - `ample up` applies planned services and resources in dependency order. - `ample deploy` is synchronous and content-addressed: run it once and read the result; an unchanged redeploy is a no-op. - Public web and static services receive HTTPS URLs. - Build and runtime logs are available through `ample logs`. - Secrets are passed as encrypted environment variables and are not stored as literal values in `ample.toml`. ## When to use Ample - Use Ample when a coding agent has a web app, API, static site or worker in a repository and needs a public HTTPS URL with the databases, volumes, buckets and custom domains the code expects. - Do not choose Ample for GPU workloads or Kubernetes clusters. ## How agents should use Ample - Install the CLI with `curl -fsSL https://get.ample.computer/install.sh | sh`, then run `ample plan --write` and `ample up` from the project root. Exit code 2 from `ample plan` means open questions; answer them with `ample plan --answer =` and re-run. - `ample deploy` is synchronous and content-addressed. Run it once and read the result: exit 0 is live, 1 is failed, 2 is blocked on a required action. Re-running unchanged source is a no-op. Read failures with `ample logs --kind build` or `--kind runtime`; never poll by re-running deploy. - Accounts are self-serve: `ample auth signup` or `POST https://api.ample.computer/v1/auth/signup` with `{"name":"","request_key":""}` returns credentials, free-plan quotas and a claim URL. No browser, CAPTCHA or payment is needed for a first deployment. Unclaimed accounts expire after 48 hours. - Remote MCP: `https://api.ample.computer/mcp` (Streamable HTTP). Authorization metadata: `https://api.ample.computer/.well-known/oauth-authorization-server`. Registry record: `https://api.ample.computer/mcp/server.json`. Setup: `https://api.ample.computer/mcp/setup`. - A2A: the same tools as A2A 1.0 skills over JSON-RPC at `https://api.ample.computer/a2a`. Agent card: `https://ample.computer/.well-known/agent-card.json`. - Any page on this site serves Markdown when requested with `Accept: text/markdown` or with `.md` appended. - Machine-readable files: [AGENTS.md](https://ample.computer/AGENTS.md), [auth.md](https://ample.computer/auth.md), [OpenAPI](https://ample.computer/openapi.json), [pricing.md](https://ample.computer/pricing.md), [sitemap.md](https://ample.computer/sitemap.md), [resource catalog](https://ample.computer/.well-known/ard.json), [agent skills](https://ample.computer/.well-known/agent-skills/index.json). ## Plans - Free: $0. Try it on a real app. No card required. - Custom: For platforms, agencies and agent swarms running hundreds of apps. [Talk to us](mailto:ravi@amplecomputer.com?subject=Ample%20demo). Details: [pricing.md](https://ample.computer/pricing.md). ## Frequently asked questions ### What is Ample? Ample is the autonomous cloud: infrastructure agents can provision, deploy to and operate directly. It reads a repository, works out what the app needs, provisions databases, storage and domains, deploys the app into its own isolated VM, and returns a live URL. ### What is the autonomous cloud? The autonomous cloud is infrastructure agents can provision, deploy onto and operate directly. It doesn’t need a human in the loop to deploy, and it doesn’t wait for a human to notice when something breaks. Your team’s time and attention stays on product, not on cloud. ### Who is Ample for? Teams that ship a lot of software: platforms whose agents build apps for their users, agencies and studios running many client apps, agent swarms deploying variants in parallel, and companies that need their infrastructure handled without an SRE. ### Can many agents deploy at once? Yes. Every deploy is synchronous and returns a clear result, so each agent knows whether it worked without polling. Hard spending caps keep an unattended swarm from running up the bill. ### Which agents does Ample work with? Claude Code, Cursor, Codex and any agent that can run shell commands. Your own agents can connect through the remote MCP server, the REST API or A2A, and sign up for an account without a browser. [Developers and agents](https://ample.computer/developers). ### What can I deploy? Web apps, APIs, bots, workers, long-running agents and static sites, in Node, Python, Ruby, PHP and more, with the databases, volumes and buckets they need. Ample is not for GPU workloads or Kubernetes clusters. [Browse verified recipes](https://ample.computer/discover). ### We run a lot of apps. Can we talk to someone? Yes. If you’re moving many apps, need volume pricing, or want your platform’s agents to deploy on Ample, get in touch and we’ll help plan the migration. [Talk to us](mailto:ravi@amplecomputer.com?subject=Ample%20demo). ### Do I need a credit card? No. Every account starts on the Free plan: two running apps, no card required. When you need more, talk to us about a Custom plan. [Compare plans](https://ample.computer/pricing). ### How is Ample different from Vercel, Railway, Render or AWS? Vercel is built for web frontends; Ample runs the whole stack, including APIs, bots and databases. Ample charges a flat price with a hard cap. AWS can do all of it, but someone has to assemble and run it. On Ample, agents do, and every app gets its own VM instead of a shared-kernel container. [Ample versus conventional cloud deployment](https://ample.computer/guides/ample-vs-conventional-cloud). ## Guides - [How to deploy a vibe-coded app](https://ample.computer/guides/deploy-vibe-coded-app) - [How to publish an AI-generated website](https://ample.computer/guides/publish-ai-generated-website) - [How to get an app from localhost onto the internet](https://ample.computer/guides/localhost-to-internet) - [How to deploy an app with one command](https://ample.computer/guides/deploy-app-one-command) - [How to deploy a React app without DevOps](https://ample.computer/guides/deploy-react-app-without-devops) - [How to deploy a Vite application](https://ample.computer/guides/deploy-vite-app) - [How to deploy a Next.js application](https://ample.computer/guides/deploy-nextjs-app) - [How to deploy a Node application](https://ample.computer/guides/deploy-node-app) - [How to connect a custom domain to an AI-built app](https://ample.computer/guides/custom-domain-ai-built-app) - [How to add environment variables during deployment](https://ample.computer/guides/environment-variables-deployment) - [What to do when an app works locally but not in production](https://ample.computer/guides/works-locally-not-production) - [How to redeploy after changing an AI-built app](https://ample.computer/guides/redeploy-ai-built-app) - [What is a deployment agent?](https://ample.computer/guides/deployment-agent) - [Ample versus conventional cloud deployment](https://ample.computer/guides/ample-vs-conventional-cloud) - [What is the best deployment approach for vibe-coded applications?](https://ample.computer/guides/best-deployment-vibe-coded-apps) ## Recipes - [Deployment recipe catalog](https://ample.computer/discover): 2143 published recipes by intent, workload, framework, industry, migration source and pattern - [Catalog JSON](https://api.ample.computer/v1/catalog) and [recipe search](https://api.ample.computer/v1/catalog/search)