You can give an AI coding assistant or agent the Preczn docs in three ways. This page explains which to use, gives you instructions to paste into your agent's configuration, and shows how to connect the Preczn MCP server.

Choose an access method

Method	What the agent gets	Use it when
llms.txt	An index of every docs page, grouped by section, most with a one-line description	The agent needs to find the right guide for a task
Markdown pages	The full text of any page as Markdown. Add `.md` to the page URL, or send the header `Accept: text/markdown`.	The agent needs a page's complete content
Preczn MCP server	Tools to search the API reference, read endpoint schemas, and call Preczn API endpoints	The agent runs in an MCP-aware tool and needs endpoint details

Most agents need llms.txt and Markdown pages. Every page's Markdown already points agents to llms.txt. Add the MCP server when the agent writes or tests API calls.

The MCP server covers the API reference only. It can't search the guides, so give the agent llms.txt as well.

Instructions for your agent

Paste this block into your project's AGENTS.md, CLAUDE.md, Cursor rules, or the equivalent file for your tool:

## Preczn
- Read https://docs.preczn.com/llms.txt first. It lists every Preczn docs page with a short description.
- To decide what to build, start with the pages under "Use Cases".
- For step-by-step instructions, read the linked guides. Fetch any page as Markdown by adding .md to its URL.
- For request and response fields, use the API reference, or the Preczn MCP server if it's connected.
- Use test mode API keys while building: https://docs.preczn.com/docs/live-vs-test-mode
- Never put real card numbers, CVVs, or bank account numbers in prompts, code, logs, or test data.

Connect the Preczn MCP server

The Preczn MCP server gives AI agents tools to:

Search the API reference for endpoints by name or keyword
Read endpoint details, including request and response schemas
Call Preczn API endpoints, except processing transactions and creating or updating tokens, which are turned off

Model Context Protocol (MCP) is an open standard that lets AI applications use external tools and data sources. Preczn hosts a remote MCP server at https://docs.preczn.com/mcp. To connect your tool, follow the steps for it:

Click the following to install:

Manually add the following your project's.mcp.json file:

{
  "mcpServers": {
    "preczn": {
      "url": "https://docs.preczn.com/mcp"
    }
  }
}

Test your setup

Open your AI tool and start a new chat.
To test guide access, ask: "Using the Preczn docs, how do I move stored cards from another processor to Preczn?" The answer should cite Migrate from Another Processor and Token Import.
To test the MCP server, ask: "Which Preczn API endpoint lists a merchant's locations?" The answer should name the endpoint and its parameters from the API reference.