# Luno OS

> Luno OS is the all-in-one platform for building robust, powerful AI agents and AI systems. Everything is real TypeScript in a real project — built for you by Luno Builder (the in-app building agent), by your own coding agent (Cursor, Claude Code, Codex, etc.), or over MCP — then deployed to a managed cloud platform with integrations, credentials, and run history. Because it's actual code rather than hidden JSON, agents can build, test, review, and maintain everything.

For AI agents and automation, use the tools below. To build without any local setup, users can use Luno Builder within the Luno OS web app — see [Luno Builder](https://os.meetluno.ai/docs/build-with-ai/platform-agent.md).

## Command line tool

Install and use the Luno OS CLI from the terminal. `luno init` scaffolds a project and `luno deploy` ships it to the platform. It also installs an `AGENTS.md` that points your coding agent to the documentation — read with `luno docs search` and `luno docs query` — so it knows how to build.

- [Install](https://www.npmjs.com/package/@meetluno/cli) — `npm install -g @meetluno/cli`
- [Agent setup guide (Markdown)](https://os.meetluno.ai/start.md) — install, auth, init, and deploy
- [CLI reference](https://os.meetluno.ai/docs/cli.md) — every command and flag

## Agent guide

`luno init` scaffolds a baseline guide into every Luno OS project as `AGENTS.md` (with a `CLAUDE.md` symlink): project layout, primitives, integrations, the deploy-first loop, and common gotchas. The CLI keeps it in sync with the installed version.

## Documentation

The documentation is the source of truth for APIs, patterns, and examples. Navigate it three ways:

- **CLI (preferred, no auth)** — `luno docs search "<query>"` to find pages, `luno docs query "cat /<path>.mdx"` to read them.
- **Index** — [Full documentation (llms.txt)](https://os.meetluno.ai/docs/llms.txt) lists every page; [llms-full.txt](https://os.meetluno.ai/docs/llms-full.txt) inlines all content.
- **URL** — append `.md` to any docs page URL for Markdown, e.g. [quickstart.md](https://os.meetluno.ai/docs/quickstart.md).

Key pages for building:

- [Workflow best practices](https://os.meetluno.ai/docs/learn/workflows/authoring-best-practices.md) — durable steps, replay, canvas rules
- [Build agents](https://os.meetluno.ai/docs/learn/agents/build-agents.md) — models, tools, memory, sandboxes
- [Company Brain](https://os.meetluno.ai/docs/learn/settings/company-brain.md) — shared, searchable organization knowledge for agents
- [Triggers overview](https://os.meetluno.ai/docs/learn/triggers/overview.md) — schedules, webhooks, polls
- [Polling and cursor state](https://os.meetluno.ai/docs/learn/triggers/polling.md) — poll filters, persisted cursor state
- [Use credentials in code](https://os.meetluno.ai/docs/learn/credentials/use-credentials.md) — declaring and resolving credentials
- [CLI reference](https://os.meetluno.ai/docs/cli.md) — every command and flag

## Integrations

Managed integrations (Luno-hosted OAuth) are coming soon. Until then, connect any HTTP API or MCP server as a custom app. `luno integrations search "{query}"` searches the live catalog.

- [Integrations (Markdown)](https://os.meetluno.ai/docs/integrations.md)
- [Custom integrations (Markdown)](https://os.meetluno.ai/docs/learn/credentials/custom-integrations.md)

<!-- Pricing is hidden from the public site for now (SHOW_PRICING in apps/web/src/lib/marketing-routes.ts).
## Pricing

Plans, features, and usage limits for building and running internal agents and AI systems.

- [Pricing (Markdown)](https://os.meetluno.ai/pricing.md)
- [Pricing (HTML)](https://os.meetluno.ai/pricing)
-->

## Models

The catalog of AI models you can use when building agents (model ids, context windows, pricing, and capabilities), refreshed on each deploy.

- [Models (Markdown)](https://os.meetluno.ai/models.md)