# Implement with AI

This documentation is also written for AI agents (Claude, ChatGPT, Cursor, Claude Code and others). You can ask them to implement CEP lookup in your project with the right context from the start.

## Buttons on every page

At the top of every page there are two buttons:

- **Copy Markdown** copies the page content as clean Markdown, ready to paste into a chat.
- **Open** opens the page straight in Claude, ChatGPT and others, already asking the agent to
  read the page and help you.

## Files for agents

| URL | What is in it |
| --- | --- |
| `/llms.txt` | An index of the documentation, with a project summary and the links. |
| `/_llms-txt/english.txt` | The whole English documentation in a single text file. |
| `/llms-full.txt` | The same documentation in Portuguese (the site's default language). |
| `/en/docs/<page>.md` | Any page as Markdown: just add `.md` to the URL. |

Point the agent at `/_llms-txt/english.txt` and it will have all the context at once.

## Ready-made prompts

Pick your environment, copy the prompt and paste it into your agent, or open it straight in Claude or ChatGPT. Each prompt includes an API summary and the link to the full documentation.

<AiPrompt framework="react" locale="en" />
<AiPrompt framework="vue" locale="en" />
<AiPrompt framework="angular" locale="en" />
<AiPrompt framework="node" locale="en" />
## Tips

- If the agent runs in the terminal (like Claude Code), paste the prompt straight into it: it reads the documentation and changes your project.
- Ask it to create the `Unicep` instance just once and turn on the [cache](/en/docs/core#cache).
- Review what it generated: check that the [errors](/en/docs/errors) are handled and
  that the [use of the providers](/en/docs/providers#using-them-responsibly) is responsible.