# agental > agental gives your agent a Linux VM over HTTP. Each sandbox is a virtual machine of its own; one request creates it, another runs a command in it, and another stops it. Every page of the docs below is Markdown. agental's skills teach an agent the whole API and move a repository into a sandbox: `npx skills add https://agental.io -g` installs them. - [Agent brief](https://agental.io/agental-brief.md): the `agental` skill, one file that teaches an agent the whole API - [Agent skills](https://agental.io/.well-known/agent-skills/index.json): every skill, for `npx skills add` - [Claude Code plugin](https://agental.io/claude-plugins.json): both skills as one plugin, for `claude plugin marketplace add` - [OpenAPI 3.1](https://agental.io/openapi.json): the full machine-readable reference ## Getting started - [Introduction](https://agental.io/docs.md): agental gives your agent a Linux VM over HTTP. - [Quickstart](https://agental.io/docs/quickstart.md): From sign-up to your first sandbox: create one, run a command in it, and stop it. - [Command line](https://agental.io/docs/cli.md): One command signs you in through your browser, installs the agental command and the agent skills, and opens a shell in a new sandbox. - [TypeScript SDK](https://agental.io/docs/typescript.md): The whole API from TypeScript and JavaScript. It waits for a machine to be running, sends a request again where that is safe, and throws every refusal as an error with the API's code. - [Python SDK](https://agental.io/docs/python.md): The whole API from Python, for plain code and for asyncio. It waits for a machine to be running, sends a request again where that is safe, and raises every refusal as an error with the API's code. - [Authentication](https://agental.io/docs/authentication.md): Every request carries your API key, as a bearer token. ## Execution - [Running commands](https://agental.io/docs/commands.md): Run a command and wait, stream its output, or start it in the background. - [GitHub Actions](https://agental.io/docs/github-actions.md): Run your GitHub Actions jobs on sandboxes by changing one line of a workflow. - [Languages and tools](https://agental.io/docs/tools.md): What every sandbox has before its first command: the languages, build tools, coding agents and assistants a task asks for most. - [Files](https://agental.io/docs/files.md): Read and write files in the sandbox work folder. - [Browser](https://agental.io/docs/browser.md): Every sandbox has a browser in it that shows as a browser on an ordinary computer does, and the command-line tools agents drive it with. Your own code can drive it from outside as well. - [Desktop](https://agental.io/docs/desktop.md): Give a running sandbox a screen: programs open windows on it, an agent reads it and clicks on it, and you watch it in a browser. - [Agent skills](https://agental.io/docs/agent-brief.md): Skills that teach an agent on your own computer to use agental: the brief, and three that move the repository it is working in into a sandbox, run agents in parallel in forks of it, and set sandboxes up with your keys and git settings. - [OpenAI Agents SDK](https://agental.io/docs/openai-agents.md): Run an agent built with the OpenAI Agents SDK in an agental sandbox. Its commands and file edits happen in the sandbox, and your code and the model's calls stay where they are. - [Schedules](https://agental.io/docs/schedules.md): Run a command every so often, or at a time of day, and have a stopped sandbox started for it and stopped again after. ## Persistence - [Sandbox lifecycle](https://agental.io/docs/lifecycle.md): Create, stop, start, archive and delete. - [Snapshots](https://agental.io/docs/snapshots.md): Keep a sandbox's disk as it is now, to roll back to or to start new sandboxes from. - [Forking](https://agental.io/docs/fork.md): Copy a sandbox to try something you may want to undo. - [Templates and secrets](https://agental.io/docs/templates.md): Make sandboxes from a template, and give every one of them the same variables and files. - [Regions](https://agental.io/docs/regions.md): Choose where in the world a sandbox runs. ## Connectivity - [SSH](https://agental.io/docs/ssh.md): Log in to a running sandbox as root, at a port of its own. - [Public URLs](https://agental.io/docs/ports.md): Give a port a public address to share what your agent built. - [Email](https://agental.io/docs/email.md): Give a sandbox an address of its own: read what is sent to it, and let it write to you. - [Webhooks](https://agental.io/docs/webhooks.md): Be told when a sandbox is made, starts, stops, fails or is deleted, and when mail arrives for one. ## Observability - [Metrics](https://agental.io/docs/metrics.md): See what a sandbox uses of its CPU, memory, disk and network. - [Usage](https://agental.io/docs/usage.md): What your sandboxes ran, and what it comes to at the list price. - [OpenTelemetry](https://agental.io/docs/opentelemetry.md): Send your sandboxes' metrics and events to Datadog, Grafana Cloud, Honeycomb or any OTLP collector. - [Tags and metadata](https://agental.io/docs/tags.md): Say what each sandbox is for, and find the ones of one run, one team or one branch. - [Billing](https://agental.io/docs/billing.md): Plans, top-ups and their bonus, and the calls that read and buy them. ## Security - [Outbound rules](https://agental.io/docs/outbound-rules.md): Decide what the sandboxes on a template may reach: everything, nothing, or a list of addresses. - [API keys](https://agental.io/docs/api-keys.md): Give each agent a key that can do only its job. - [Retries and idempotency](https://agental.io/docs/retries.md): Agents crash, time out and try again. Retrying is always safe. - [Your account and data](https://agental.io/docs/account.md): What agental keeps about your account and you, how long, how to take it away, and how to close the account. - [Two-factor sign-in](https://agental.io/docs/two-factor.md): A second step after every sign-in: six digits from an app on your phone, a backup code, or a passkey. ## Reference - [API endpoints](https://agental.io/docs/endpoints.md): The calls for sandboxes, templates, snapshots, email, metrics, telemetry, usage, billing and webhooks. The machine-readable version of everything is OpenAPI 3.1. - [Errors](https://agental.io/docs/errors.md): The errors a sandbox call answers with, what each means, and what to do next. - [Limits](https://agental.io/docs/limits.md): Every number in one place.