Skip to content
SteadyLinkDocs
Dashboard

Set up SteadyLink with an AI assistant

Point ChatGPT, Claude, Cursor or any other AI assistant at the SteadyLink docs so it can walk you through setup or write the integration for you.

On this page

AI assistants are good at setup work: writing an upload script, wiring SteadyLink into your app, or configuring rclone. They do it well only when they read the real docs instead of guessing. SteadyLink publishes its documentation in two plain-text files made for that.

FileWhat it holdsUse it when
llms.txtA short index: what SteadyLink is, the key integration facts, and a link to every docs page.The assistant can browse the web and will open the pages it needs.
llms-full.txtThe full text of every docs page in one file, including every API endpoint, SDK method and code example.You want the assistant to have everything at once, or it can't follow links.

Both files are rebuilt from the docs on every release, so they always match this site.

Ask an assistant to set things up#

  1. Give it the docs

    In ChatGPT, Claude, Gemini or another assistant that can browse, paste the link https://steadylink.io/llms-full.txt at the start of your message. If it can't open links, download llms-full.txt and attach the file instead.

  2. Say what you want to build

    Be specific about your stack and the result you want. For example:

    Text
    Read https://steadylink.io/llms-full.txt first. Then help me write a Node.js
script that uploads every PDF in ./catalogs to a SteadyLink bucket called
"catalogs" and prints each file's permanent link. When I run it again, it
should replace files that already exist so their links stay the same.

  3. Keep your key out of the chat

    The assistant only needs to know that a key exists. Create the key yourself in Dashboard > Developers, store it in an environment variable such as STEADYLINK_API_KEY, and have the code read it from there. Never paste a full slk_ key or S3 secret into a chat.

In code editors#

Cursor, VS Code with GitHub Copilot, Windsurf, Claude Code and similar tools can read the same file.

  • Cursor: add https://steadylink.io/llms-full.txt under Settings > Indexing & Docs, then mention it with @ in chat.
  • Claude Code, Codex and other terminal agents: start your request with "Read https://steadylink.io/llms-full.txt first", or save the file into your project and reference it.
  • Any editor: save llms-full.txt next to your code so the assistant picks it up as project context.

Good things to ask for#

  • "Set up rclone so I can sync a folder to SteadyLink" (see also Connect rclone in one command).
  • "Add an upload button to my Next.js app that sends files straight to SteadyLink without exposing my API key" (see Browser uploads).
  • "Write a script that replaces our price list PDF every Monday so the link on our site never changes" (see Upload and replace files).
  • "Serve our product images through next/image with resizing" (see Images in Next.js).

If an assistant gives you an endpoint, parameter or SDK method you can't find in these docs, it is guessing. Ask it to quote the part of llms-full.txt it used, or check the REST API reference.