cd /news/ai-products/show-hn-using-a-diffusion-model-to-w… · home topics ai-products article
[ARTICLE · art-133674] src=cortee.ai ↗ pub= topic=ai-products verified=true sentiment=↑ positive

Show HN: Using a diffusion model to write docs quickly

Cortee launched a documentation tool that uses a diffusion model to apply edits to live documentation pages in under one second, with typical edit latency of around 0.7 seconds in internal tests. The tool takes a short written note describing what changed, proposes a small edit tied to that fact, and leaves the user in control of what remains on the page, with documents stored in the EU on AWS in Ireland and sign-in, search, and generation running in the United States. Cortee states it does not train on or sell customer data, and that access is separate per organization, restricted, and logged.

by read2 min views2 publishedSep 18, 2026
Show HN: Using a diffusion model to write docs quickly
Image: source

with minimal effort.

When you tell Cortee what changed, it updates the live page in under a second<sup>1</sup>.

From a written note to a reviewed change #

Write what you know. Cortee applies the edit. You remain in control.

  1. Write what you knowThink of it as your notepad. Use a few words to describe what changed.

  2. Cortee applies the editThe edit lands on the live page, so you can see what changed right away.

  3. You remain in controlEvery paragraph is backed by the facts you wrote, so you can see where any line came from.

Why Cortee #

A narrower approach to generated documentation #

A wiki leaves the writing to people, so trust stays high and updates stay slow. A markdown repository can rewrite entire pages, then leave changes that are hard to follow. Cortee takes a narrower path. From the fact you provide, it proposes a small change tied to that fact, and you decide what remains on the page.

AI documentation that stays current · Docs without Git · Compare

Your data stays yours. #

For more information, visit our Privacy Policy and Data Processing Agreement.

What we never do

  • We do not train on your data.
  • We do not sell your data.

Where it lives

  • Documents are stored in the EU on AWS in Ireland, including history and logs.
  • Sign-in, search, and generation currently run in the United States.

Access and control

  • Access is separate for each organization.
  • Access is restricted and logged.
  • You can export or delete your data on request.

When a process changes, the documentation should change with it. #

<sup>1</sup> Typical edit latency is around 0.7s in internal tests. Performance varies by prompt complexity, network, and system load.

── more in #ai-products 4 stories · sorted by recency
── more on @cortee 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/show-hn-using-a-diff…] indexed:0 read:2min 2026-09-18 ·