# Use with AI (https://docs.predictorsdk.com/ai-tools)

Hand these docs to ChatGPT, Claude, Cursor or any coding agent as clean Markdown.

Every page here is also published as plain Markdown, so an AI tool can read the
docs without the website around them.

## One page [#one-page]

Under each page's title:

- **Copy Markdown** copies the page as Markdown, ready to paste into a chat.
- **Open** sends the page to ChatGPT, Claude or Cursor, or shows its Markdown.

The Markdown copy lives at the page's address with `.md` on the end:

```txt
https://docs.predictorsdk.com/quickstart.md
https://docs.predictorsdk.com/api-reference/getMarkets.md
https://docs.predictorsdk.com/changelog.md
```

A tool that asks for Markdown (`Accept: text/markdown`) gets it from the
normal address too.

## The whole site [#the-whole-site]

| File | What it holds |
|---|---|
| [llms.txt](https://docs.predictorsdk.com/llms.txt) | A short guide to the API and a link to every page's Markdown copy. Start here. |
| [llms-full.txt](https://docs.predictorsdk.com/llms-full.txt) | Every page and the full changelog in one file, for tools that read a whole document at once. |
| [openapi.yaml](https://docs.predictorsdk.com/openapi.yaml) | The OpenAPI spec: every endpoint, parameter and response schema. The SDKs are generated from it. |

## In a coding agent [#in-a-coding-agent]

Give the agent `llms.txt` along with the task. For example:

```txt
Read https://docs.predictorsdk.com/llms.txt. Then write a TypeScript script
that lists today's NFL games and prints each game's Kalshi and Polymarket
prices for the full-game winner.
```

The agent can follow the links in `llms.txt` to the pages it needs. If it
writes code against the SDKs, point it at the [SDK pages](https://docs.predictorsdk.com/sdk/installation)
as well: their tables list every method and field.