KDP AI Book Maker API (via MCP)

Programmatic access to your books, chapters, metadata, and exports — built on the Model Context Protocol. No bespoke SDK. No static API keys.

Open KDP AI Book Maker See how it works

4.9/5 by 12,847 authors

Cancel anytime

12,000+ authors publishing

KDP-ready PDFs in minutes

100% commercial rights

How it works

One endpoint, every tool

POST to https://kdpaibookmaker.com/mcp with a JSON-RPC 2.0 payload. Same schema as every other MCP server.

Read your library

list_books, get_book, get_profile — reads run without approval prompts, ideal for dashboards and reporting.

Write with confirmation

Chapter drafting, metadata refreshes, and exports all follow MCP's approval hints. Great for pipelines that want a human in the loop.

OAuth 2.1 secured

No static API keys. Every caller registers dynamically (RFC 7591) and users grant per-client consent.

What's included

JSON-RPC 2.0 over HTTPS
OAuth 2.1 + dynamic client registration
Self-describing tool schemas
Works with every MCP SDK
Same credits as the app
Ideal for pipelines & dashboards

Frequently asked questions

Is there a REST API too?

We ship MCP because it covers 95% of what a REST API would, plus tool schemas, auth, and streaming out of the box. Any HTTP client can call MCP — it's just JSON-RPC over POST.

What are the rate limits?

Tool calls draw from your plan's book credits. Reads are unmetered on paid plans. Writes and exports use one credit each.

Can I use this from a script?

Yes. Any MCP SDK (Python, TypeScript, Rust, Go) can connect. You can also POST raw JSON-RPC from curl once you have an OAuth token.

Where's the tool reference?

The full tool list plus JSON schemas is discoverable via the MCP endpoint itself — call the standard 'tools/list' method after auth.

Related guides

Keep exploring AI book creation and Amazon KDP publishing.