Launched: briefroom — share AI-generated HTML in 30 seconds, pull client feedback back as LLM-ready Markdown

Hey IH! 👋

Today I launched briefroom (https://briefroom.net) — solo-built in

6 weeks, from Japan.

The itch: I do client work with AI coding agents (Claude Code, Codex).

They generate HTML proposals and mockups in seconds — but then the workflow

collapses:

- Hosting a static HTML draft just to show a client = Vercel overkill

- Feedback comes back as vague emails ("can you fix the top part?")

- I manually translate every comment back into prompts

Generation takes 30 seconds; sharing and collecting feedback takes 30 minutes.

What briefroom does:

1. Drop an HTML folder (or npx @briefroom/cli deploy ./) → share URL in

30s. No signup for the first deploy.

2. Clients click/tap any element on the live page and comment right there.

No account needed, mobile-first (decision-makers review on phones).

3. briefroom feedback pull returns every comment as structured,

LLM-ready Markdown → paste into your agent → fix everything → redeploy

to the same URL. Loop closed.

It's llms.txt-compliant, so agents can learn the whole workflow themselves.

Business model: Free / Pro at ¥690 (~$5/mo) / Founders Edition —

one-time ¥4,800 (~$30) for lifetime Pro, capped at 50 seats as a launch

experiment. Targeting Japanese freelancers & consultants first (underserved,

non-English-first market), English UI is fully supported.

Some honest numbers & bets:

- 6 weeks from first commit to launch, built almost entirely with Claude Code

- Infra cost is ~$30/mo at current scale (R2 zero-egress does heavy lifting)

- Biggest risk I'm watching: abuse (it's "host arbitrary HTML" as a service —

isolated delivery domain, forced CSP, malware scanning from day one)

Also on Product Hunt today: https://www.producthunt.com/products/briefroom

Ask me anything — especially curious how others here handle the

client-feedback round-trip in AI-assisted work. 🙏