> ## Documentation Index
> Fetch the complete documentation index at: https://docs.reactor.inc/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Reactor hosts multiple models, each with its own connect slug (modelName) and command/event schema. The catalog of every model — slug, typed SDK package, and links to its schema — is at /model-api-reference/overview. Some models expose one slug per experience (e.g. HappyOyster); always take the slug from the model's own pages, never guess it.
> Fastest path to a working app: `npx create-reactor-app my-app --model=<slug>` scaffolds a complete app with secure auth wired up. Typed TypeScript SDKs are published as @reactor-models/<model>; Python uses the base reactor-sdk package.
> Auth: exchange an API key (rk_...) for a JWT via POST https://api.reactor.inc/tokens from your server. Never put the API key in client-side code.
> Append .md to any docs URL for clean Markdown. Search these docs via the MCP server at https://docs.reactor.inc/mcp.

# Build with AI agents

> Point a coding agent at Reactor's machine-readable docs, the MCP server, and a bootstrap prompt that gets an agent from zero to a running app

Everything on this site is built to be read by coding agents as well as people. This page collects
the machine-readable surfaces in one place, so you can hand your agent one link and let it work.

## Machine-readable docs

* **[llms.txt](https://docs.reactor.inc/llms.txt)**: the documentation index. Every model with its
  connect slug and typed SDK package, plus the core guides. The standard first fetch for an agent.
* **[llms-full.txt](https://docs.reactor.inc/llms-full.txt)**: the entire documentation in one
  file. Large; prefer the index or a specific page.
* **Any page as Markdown**: append `.md` to any docs URL (for example
  [/quickstart.md](https://docs.reactor.inc/quickstart.md)), or send an `Accept: text/markdown`
  header to the normal URL.

## Docs search over MCP

An MCP server at `https://docs.reactor.inc/mcp` exposes docs search and a read-only filesystem
view of every page.

```shell Claude Code theme={"theme":{"light":"github-light","dark":"github-dark-high-contrast"}}
claude mcp add --transport http reactor-docs https://docs.reactor.inc/mcp
```

In Cursor or VS Code, use the contextual menu on any docs page (the buttons next to the page
title) for one-click MCP install.

## Bootstrap prompt

Paste this into your agent to go from nothing to a running app:

```text Bootstrap prompt theme={"theme":{"light":"github-light","dark":"github-dark-high-contrast"}}
Build me an app on Reactor (https://docs.reactor.inc).
1. Fetch https://docs.reactor.inc/llms.txt and pick a model from the Models
   section. Each entry lists the model's connect slug and typed SDK package.
2. Scaffold with: npx create-reactor-app my-app --model=<slug>
   (run it without --model to list available models).
3. I'll put my API key in .env as REACTOR_API_KEY. Never put the key in
   client-side code; the scaffolded app already mints short-lived tokens
   server-side via POST https://api.reactor.inc/tokens.
4. Read the model's schema page (linked from llms.txt) before sending
   commands, and take model slugs from the docs, never from memory.
```

Get an API key from the [Reactor dashboard](https://reactor.inc/dashboard). For token scoping,
see [Authentication](/authentication).

## What agents should know

* Every model has its own command and event schema. The
  [model catalog](/model-api-reference/overview) links each model's schema, prompt guide, and
  tutorial.
* Model slugs are not guessable from names: some models are namespaced differently, and some expose
  one slug per experience. Always read the slug off the model's own pages.
* `create-reactor-app` scaffolds a complete app for any model in the catalog, with the SDK
  installed and a secure token route wired up.