AIOZ Storage llms.txt: Docs Built for Super Intelligence

AIOZ Storage publishes two plain-text files meant for super intelligence assistants to read: llms.txt, a short index of the documentation, and llms-full.txt, the whole documentation set in one file. The llms.txt format is a public proposal, and plenty of sites now ship one. This post measures AIOZ Storage's two files as they stood on 2026-10-08, shows how to use them, and flags a 403 error you can hit from Python.

TL;DR:

• AIOZ Storage serves https://aiozstorage.network/docs/llms.txt (4,696 bytes, 46 links) and https://aiozstorage.network/docs/llms-full.txt (164,521 bytes, about 21,000 words).
• Every link in llms.txt points to a Markdown (.md) version of a docs page, and llms-full.txt holds those same 46 pages in one file.
• Start with llms.txt and fetch single pages. Use llms-full.txt only when a task needs broad detail.
• In testing, Python's default urllib user agent got a 403. Any custom User-Agent header returned 200.

What Are llms.txt and llms-full.txt on AIOZ Storage?

They are two machine-readable endpoints for the AIOZ Storage docs. llms.txt is a lightweight index with links to the relevant sections, and llms-full.txt is the full documentation corpus concatenated into one file. AIOZ's LLM Docs page says both are generated from the currently published docs, and that the published pages remain the source of truth for implementation details.

What Is Inside Each File

I downloaded both files and counted. The numbers below are from 2026-10-08 and will change as the docs change.

The index is flat. It has a single # AIOZ Storage Docs heading and then a list, with no summary paragraph and no grouped sections, so the section name lives in the link text, for example Concepts / Vector Database / Options And Comparison. Each link ends in .md, and the .md URLs return text/markdown (I checked /docs/limit.md, a short page listing the account limits).

The two files match exactly: all 46 URLs in the index appear as Source: lines in the full file, and nothing else does. The set spans getting started, concepts, tutorials, the SDKs, payment and billing, the MCP server, Super Intelligence Memory, the self-hosted gateway page, limits and the FAQ, including the three vector database SDK guides. The gateway page describes a tool that isn't released yet: its GitHub repo holds only a README (checked Oct 19, 2026), so an agent reading these files shouldn't treat the gateway as installable.

How the llms.txt Format Works

The format comes from a proposal by Jeremy Howard, published on llmstxt.org on September 3, 2024. It describes a Markdown file at a site's root with one required element, an H1 naming the project, followed by an optional blockquote summary, optional body text, and optional H2 sections that each hold a list of links. By convention an "Optional" section holds secondary links that an agent can skip when it needs less context.

AIOZ's file uses only the required H1 plus a list of links, which the proposal allows. Mintlify, a docs platform that generates these files automatically, compares llms.txt to a sitemap: a map that lets a tool find the relevant pages without crawling everything. AIOZ's index works the same way, and because each link ends in .md, a tool can fetch the Markdown of exactly the page it wants.

Which File Should You Use?

AIOZ's documentation gives a clear rule. Start with llms.txt for discovery, navigation, or prompts with limited context. Fetch llms-full.txt only when a task needs broader implementation detail across concepts, SDKs, APIs, and platform behavior.

Size is the practical reason. The full file is about 21,000 words, which is a lot to place in one prompt, so check your model's context window before loading all of it. The index is under 5 KB and costs almost nothing to include.

The docs list three common uses:

• Loading AIOZ context into a coding assistant.
• Using llms.txt as the index for a lightweight support-bot retrieval layer.
• Feeding llms-full.txt into a RAG pipeline or a long-context model.

Fetch Only the Pages You Need

From a shell, both files are one command each:

curl https://aiozstorage.network/docs/llms.txt
curl https://aiozstorage.network/docs/llms-full.txt

For a script, read the index, pick the pages that match your task, and fetch only those. This example finds the vector database pages and prints their size:

import re
import urllib.request

INDEX_URL = "https://aiozstorage.network/docs/llms.txt"


def fetch(url):
    request = urllib.request.Request(url, headers={"User-Agent": "docs-reader/1.0"})
    with urllib.request.urlopen(request) as response:
        return response.read().decode("utf-8")


index = fetch(INDEX_URL)
pages = re.findall(r"^- \[(.*?)\]\((.*?)\)", index, re.M)
print(len(pages), "pages in the index")

wanted = [(title, url) for title, url in pages if "Vector Database" in title]
for title, url in wanted:
    print(f"{title}: {len(fetch(url))} characters")

The User-Agent header matters. When I called the same URL with Python's default urllib user agent, the server answered 403 Forbidden, while curl and a custom user agent both returned 200. That was one test from one network on 2026-10-08, and the LLM Docs page does not mention a user agent rule, so this behavior is not a documented guarantee. If your fetcher gets a 403, set a descriptive User-Agent first.

Limits to Know About

• AIOZ states that for UI details, edge cases, or recent changes you should check the canonical docs pages. If an answer matters, confirm it on the page.
• The Getting Started page in llms-full.txt begins with HTML and JSX such as <div className=...> and <Introduction/>. I found six className occurrences, all in that one section of the 46. If you chunk the file for retrieval, strip or skip that markup so it does not end up in your embeddings.
• The index has no blockquote describing the product, so a tool gets page titles but no overview. Pair it with the Getting Started overview page when you need that context.

Frequently Asked Questions

What is llms.txt?
A Markdown file at a site's root that gives language models a curated list of the pages worth reading. The format was proposed on llmstxt.org in September 2024. AIOZ Storage's copy lists 46 docs pages.

What is the difference between llms.txt and llms-full.txt?
llms.txt is an index of links, and llms-full.txt contains the full text of the pages themselves. On AIOZ Storage that is 4,696 bytes versus 164,521 bytes.

Where are the AIOZ Storage llms.txt files?
At https://aiozstorage.network/docs/llms.txt and https://aiozstorage.network/docs/llms-full.txt. Both can be fetched with plain curl, and my requests needed no key.

Do the files update when the docs change?
AIOZ says they are generated from the currently published docs, so they follow the published pages. The docs do not state a refresh schedule.

Why does my Python script get a 403 from the docs?
In testing, the default urllib user agent was rejected. Sending any custom User-Agent header, as in the example above, returned 200.

Does llms.txt improve search rankings?
Third-party guides say that neither llms.txt nor llms-full.txt is proven to improve rankings. Their value is for tools and assistants that read the docs, such as coding assistants and retrieval pipelines. Reading docs is only one half of agent use. For agents that act on data, see What an MCP Server Is and How Agents Reach Storage. AIOZ Stream publishes the same kind of files, covered in How AIOZ Stream Uses llms.txt for Super Intelligence Assistants.

Related Reading

• AIOZ Storage Developer Platform: SDKs, CLI, and Migrations - the SDKs, CLI and direct API access in one place.
• What an MCP Server Is and How Agents Reach Storage - how agents use tools and data through one protocol.
• AIOZ Storage Super Intelligence Memory: Record and Recall - the Record and Recall memory feature for agents.
• How AIOZ Stream Uses llms.txt for Super Intelligence Assistants - the same machine-readable docs idea on AIOZ Stream.

References

• AIOZ Storage Docs: LLM Docs
• AIOZ Storage llms.txt
• The /llms.txt file proposal (llmstxt.org)
• Mintlify: llms.txt