AI Version (llms.txt)

ProxyHat publishes machine-readable copies of this documentation so AI assistants can read and reason about the API directly. If you use Claude, Cursor, Codex, or any other LLM-powered tool to build your integration, hand it these files and it will have the full, accurate picture of the API without guessing.

Give your AI the whole API

One click copies the complete documentation plus a short primer that tells your assistant how to use it — like installing a skill. Paste it into a new chat and the AI knows the whole ProxyHat API.

Works with Claude, ChatGPT, Cursor, Codex, Gemini, and any other assistant.

Two artifacts, one convention. ProxyHat follows the llmstxt.org convention. /llms.txt is a concise, curated index of links to every documentation page. /llms-full.txt is the entire documentation flattened into a single self-contained Markdown file — everything an assistant needs in one paste.

The two files

Link index — a short Markdown file listing every docs page with a one-line summary and its URL. Small enough to keep in an assistant's context permanently; the model fetches individual pages on demand.

https://docs.proxyhat.com/llms.txt

Full documentation — every page concatenated into one Markdown document. No links to follow, no fetching required: paste it once and the assistant knows the whole API.

https://docs.proxyhat.com/llms-full.txt

How to use it with your tool

Two approaches work everywhere:

Paste it. Click "Copy for AI" above (or open https://docs.proxyhat.com/llms-full.txt and copy the whole file), then paste it into your assistant's chat or a context/rules file. Best when your tool can't reach the internet.
Point at the URL. Give your tool the https://docs.proxyhat.com/llms.txt or https://docs.proxyhat.com/llms-full.txt URL and let it fetch. Best for tools with web or file-fetch access.

Tool-specific notes

Claude / Claude Code — paste the contents of llms-full.txt into the conversation, or add the URL to your project's CLAUDE.md so every session picks it up. Claude Code can also fetch the URL directly with its web tools.
Codex — drop the URL or the pasted contents into your AGENTS.md (or the prompt) so the agent references it while writing your integration.
Cursor — add https://docs.proxyhat.com/llms-full.txt under Settings → Indexing & Docs (Add new doc), or paste the file into a .cursorrules / project rule.
Kimi — paste the contents of llms-full.txt into the chat, or share the URL and ask it to read the file first.
OpenCode — reference the URL in your agent config / rules file, or paste the full file into context.
Any other LLM — the rule is the same: use the "Copy for AI" button, paste the contents of llms-full.txt into your assistant, or point your tool at the URLs above.

What the AI will learn

From these files an assistant gets everything it needs to write a working integration, including:

Base URL — every request goes to https://api.proxyhat.com/v1.
Authentication — how to obtain an API key and send it as Authorization: Bearer <key>.
The sub-user model — how sub-users, groups, and presets map onto proxy access.
Proxy connection strings — the username grammar (country, session, rotation, and AI-filter targeting) and how to build or fetch a ready-to-use connection string.
The reseller flow — provisioning and managing proxies on behalf of your own customers.
Response shapes — which endpoints use the standard {success, payload, meta, errors, description} envelope and which return flat objects, with real field names throughout.

Always current. Both files are generated from the same source as the pages you're reading now, so an assistant working from them stays in sync with the live API. When in doubt, prefer llms-full.txt — it is complete and self-contained.