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

# Overview

> Connect any AI client to Tolstoy and run your shoppable video workspace from chat — via the Model Context Protocol (MCP).

Tolstoy exposes its platform to AI clients over the **Model Context Protocol (MCP)** — the open standard for connecting assistants like Claude, ChatGPT, and Cursor to external tools. Connect once, then run your **shoppable video** workspace right from chat: generate content, manage your media library, build and publish shoppable widgets, tag products, and track ad performance.

There are **two remote MCP servers**, both served over Streamable HTTP and secured with OAuth 2.1 + PKCE. There are no API keys to paste — your client runs the OAuth sign-in on first connect, and the connection is bound to the Tolstoy workspace you authorize with.

| Server              | Endpoint                                         | What it does                                                                                                                                      |
| ------------------- | ------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Tolstoy Library** | `https://apilb.gotolstoy.com/mcp/v1/library/mcp` | Run your shoppable video workspace — media library, shoppable widgets (Stories, Carousel, Spotlight), product tagging, multi-store, and Meta ads. |
| **Tolstoy Studio**  | `https://apilb.gotolstoy.com/mcp/v1/mcp`         | Generate and iterate on marketing videos and images with Tolstoy Studio's AI creative director.                                                   |

<CardGroup cols={2}>
  <Card title="Connect a client" icon="plug" href="/mcp/connect">
    Step-by-step setup for ChatGPT, Claude, Cursor, and more.
  </Card>

  <Card title="Tolstoy Library tools" icon="photo-film" href="/mcp/library">
    Assets, shoppable widgets, products, stores, and ads.
  </Card>

  <Card title="Tolstoy Studio tools" icon="wand-magic-sparkles" href="/mcp/studio">
    Generate and iterate on videos and images.
  </Card>

  <Card title="MCP Registry" icon="box-archive" href="https://registry.modelcontextprotocol.io">
    Published as io.github.GoTolstoy/library and /studio.
  </Card>
</CardGroup>

## Authentication

OAuth 2.1 with PKCE, backed by Amazon Cognito. Clients discover the flow via RFC 9728 protected-resource metadata at each server's `/.well-known` endpoints — there is nothing to configure beyond adding the server URL and signing in.

<Note>
  Each connection operates on the Tolstoy workspace you sign in with. Accounts with multiple stores can target any connected store from the widget, product, and ads tools.
</Note>

## App-aware clients

In clients that support interactive MCP apps (ChatGPT, Claude), tool results render as inline views — a shoppable-widget card grid, asset previews, and widget detail with the actual videos a widget shows. Plain clients receive the same data as text, so every tool works everywhere.

## Get started

The fastest path is the published **Tolstoy Library** app in ChatGPT — search "Tolstoy Library" and click Connect. For every other client, see [Connect a client](/mcp/connect). Per-client setup is also available in the Tolstoy platform under **Settings → MCP**.