# ORB Cloud

**Deploy your AI agent. We run it.**

ORB Cloud is a deployment platform for AI agents. Push your code, we give it a computer — full Linux, bash, packages, network. Your agent runs on our infrastructure. You pay a fraction of what a VM costs.

## How It Works

1. You write your agent (Python, Node.js, Go, Rust — anything that runs on Linux)
2. You write an `orb.toml` config file (tells ORB what to run, what to install, where your LLM provider lives)
3. You deploy via the API (create computer → upload config → build → deploy)
4. Your agent is live at `https://{id}.orbcloud.dev`

ORB handles everything else — infrastructure, scaling, and cost optimization. Your agent just runs.

## Where to start

**Get going:**
- [Why ORB](why-orb.md) — Who it's for, when to use it, when not to, and how it compares to VPS / Docker / Lambda / E2B.
- [Getting Started](getting-started.md) — Step-by-step walkthrough: register, create a computer, write `orb.toml`, build, deploy.

**Reference:**
- [orb.toml Reference](config-reference.md) — Every section of the config file. Backends for Anthropic, OpenAI, Google, OpenRouter, custom — and the schema for everything else.
- [API Reference](api-reference.md) — All endpoints: computers, config, build, deploy, files, terminal, webhooks, usage.

**Help and cost:**
- [Pricing](pricing.md) — Plans, GB-hour rates, calculator.
- [Troubleshooting](troubleshooting.md) — Real problems and how to fix them.

**For AI agents:**
- [llm.txt](/llm.txt) — Entire ORB Cloud reference on a single page. The fastest way for an agent to learn how to deploy: `curl https://docs.orbcloud.dev/llm.txt`

## Base URL

```
https://api.orbcloud.dev
```

## Authentication

ORB API keys are self-serve. Register with just an email and the response contains a working key — no waitlist, no payment, no human approval required.

```bash
curl -X POST https://api.orbcloud.dev/api/v1/auth/register \
  -H 'Content-Type: application/json' \
  -d '{"email":"you@example.com"}'
```

Use the returned `api_key` on every subsequent call:

```
Authorization: Bearer <api_key>
```

See [Getting Started](getting-started.md) for the full register → deploy walkthrough.

---

## For AI agents (single-prompt deploy)

If you're driving an AI agent that needs to deploy itself to ORB, paste the prompt below into your agent's context. It bootstraps the entire flow — register, create, build, deploy — using only the documented HTTP API.

The fully expanded reference (orb.toml schema, every endpoint, all examples) is at [llm.txt](/llm.txt) — one curl, one page, designed for AI consumption.

```
You have access to ORB Cloud — a deployment platform for AI agents.

API base: https://api.orbcloud.dev

To deploy your agent:
0. If you don't already have an ORB API key, register one yourself — it's
   self-serve, instant, free, no waitlist, no payment, no human approval:
     curl -X POST https://api.orbcloud.dev/api/v1/auth/register \
       -H 'Content-Type: application/json' \
       -d '{"email":"agent+<unique>@yourdomain.com"}'
   The response contains api_key — use it as `Authorization: Bearer <api_key>`
   for every call below. Save it; the email is the only way to recover it.
1. Read the full getting started guide: curl https://docs.orbcloud.dev/getting-started?format=md
2. Read how to write orb.toml: curl https://docs.orbcloud.dev/config-reference?format=md
3. Create a computer: POST /v1/computers with {"name": "my-agent", "runtime_mb": 2048, "disk_mb": 4096}
4. Write an orb.toml for your codebase (see config reference for format and examples)
5. Upload it: POST /v1/computers/{id}/config with Content-Type: application/toml
6. Build: POST /v1/computers/{id}/build (clones repo, installs deps — use curl -m 600)
7. Deploy: POST /v1/computers/{id}/agents with {}
8. Your agent is live at https://{first-8-chars-of-id}.orbcloud.dev

The orb.toml tells ORB:
- [agent] — what to run (language, entry point)
- [source] — where your code is (git repo)
- [build] — how to install dependencies
- [llm] — your LLM endpoint (base_url for proxy)
- [ports] — what ports to expose

Read the config reference for the full spec and examples for every provider.
```