Start here when an LLM needs to discover Qorven’s docs. It’s a flat newline-delimited list of URL — description entries, matching the convention adopted by llmstxt.org.

​
Canonical URL

https://docs.qorven.ai/llms.txt
Mirrored on every Qorven install that serves docs locally.

​
Format

# Qorven documentation

Machine-readable index of every documentation page. Each line:
`<URL> — <description>`.

## Pages
https://docs.qorven.ai/getting-started/install — One-line install...
https://docs.qorven.ai/getting-started/first-chat — Open the web UI...
https://docs.qorven.ai/architecture/overview — The 5-minute tour...
...

​
How to use it

1

Fetch once per session

An agent chatting with a user does one GET to the index on startup, caches the result.
2

Scan for relevance

Match the user’s question against the descriptions. Pick the top N (3–5) most-likely relevant URLs.
3

Fetch the markdown

Swap .mdx → .md on any doc URL, or append .md to the path, to get the plain markdown source. Faster to parse than the HTML.
4

Quote it back

In your reply, link the source URL so the user can read the full page.

​
Example

import requests

index = requests.get("https://docs.qorven.ai/llms.txt").text.splitlines()
# find any line mentioning "encryption key"
hits = [l for l in index if "encryption" in l.lower()]
for l in hits[:3]:
    url = l.split(" — ")[0]
    md = requests.get(url + ".md").text
    # feed md into your LLM

​
Refresh cadence

Regenerated whenever the docs build runs. Typical lag: minutes after a PR merges.

​
Where next

Agent guide

Full conventions for LLM consumption.

Markdown source

How to fetch the raw .md for any doc.

docs_search tool

For Qors: semantic search at runtime.