AI & Agent Access | HyperFlow Docs

These docs are published in machine-readable form as well as human-readable form, so an AI assistant or coding agent can read them directly instead of scraping the rendered site.

Connect the MCP server

The fastest path for a coding agent. The server exposes a searchDocs tool that returns doc passages with their source URLs.

Endpoint: https://docs.hyperflow.fun/_mcp/server (streamable HTTP, no auth)

claude mcp add --transport http hyperflow-docs https://docs.hyperflow.fun/_mcp/server

Once connected, ask the agent things like “how do I sign a HyperFlow perps order?” and it will search these pages and cite them.

Fetch the docs as text

If you would rather pull the content yourself — for a RAG index, an eval, or a one-off prompt:

URL	What you get
`/llms.txt`	An index of every page: title, URL, one-line description. Follows the llmstxt.org spec
`/llms-full.txt`	The entire documentation concatenated as one Markdown document
`<page URL>.md`	That one page as clean Markdown
`<section URL>/llms.txt`	An index scoped to one section

For example, /developers/perps-api/authentication.md is that page alone, and /developers/perps-api/llms.txt indexes just the API reference.

# The whole site, as one Markdown file
curl -s https://docs.hyperflow.fun/llms-full.txt
# Just the API reference pages
curl -s https://docs.hyperflow.fun/developers/perps-api/llms.txt

Every page also carries Copy page, View as Markdown, and Open in Claude / ChatGPT actions in its header.

What an agent should know

Worth putting in your agent’s system prompt when it works against these docs:

Configuration is live, not literal. Fee rates, margin ratios, order limits, minimums, and the HyperCore deposit address are served by GET /markets and GET /config. Any number written in these docs is an illustration of the shape, not a value to hardcode.

Ask a question

For anything the docs do not answer, the developer channel on Discord is the fastest route to a human.