Launched Stoopwriter — AI local SEO content for real estate agents. $39 MRR (1 customer, me). Going for 20 founding seats in 30 days.
Hey IH. Just shipped Stoopwriter, an AI content generator tuned specifically for local real estate agents.
The pitch:
Most AI writers produce content that sounds like it was written from Iowa about a city in Florida. For local SEO (which is 80% of how realtors get organic leads), that's a problem. Stoopwriter takes the agent's city, neighborhoods, and target keywords and outputs content that names specific transit lines, price points, and architectural styles instead of "stunning home in a vibrant community" boilerplate.
Five real samples behind a locked demo here: https://stoopwriter.com/samples?utm_source=indiehackers&utm_medium=launch&utm_campaign=launch1
The stack:
- Next.js 14 App Router on Vercel
- Anthropic Claude (Sonnet for long-form, Haiku for shorts)
- Upstash Redis for per-user usage counters
- Stripe for billing
- Resend for transactional email
- PostHog for product analytics
- llm-eval harness with Claude-as-judge for prompt regressions
What's working:
- End-to-end purchase → webhook → auto-provision → welcome email works in under 60 seconds
- Eval rubric (5 dimensions) lets me catch prompt regressions before deploying. Latest run: 19.7/25 average across 18 generations, up from 18.6 after three targeted prompt fixes
- Founding tier ($39/mo, first 20 Vancouver agents, locked for life) gives me both early revenue and a real reason for cold outreach
What's not yet:
- Cold outreach reply rate too early to measure: 24 emails sent over the last 48h, still in the 24-72h window
- Conversion from /samples to paid is untested
- No /blog yet for long-tail SEO. Just shipped 6 long-form posts targeting realtor-content keywords; will know in 4-12 weeks if it ranks
Asks:
- If you've built B2B vertical AI tools, what early-customer-acquisition surprises did you hit?
- For prompt evals, anyone else running LLM-as-judge with the same model that generated the output? Worried about bias.
- Roast my pricing. $39 founding / $79 Starter / $149 Pro / $299 Agency. Too low? Too high?
Happy to share the eval harness code if useful.
This feels stronger than a normal AI writing tool because the wedge is not really “write me content.” It is helping agents publish content that sounds locally credible instead of generic.
The examples you gave around neighborhoods, transit lines, price points, and architectural styles are the real differentiator. That is the part I’d push harder on the landing page, because agents do not just want more blog posts. They want to look like the obvious local expert when a buyer or seller searches.
One thing I’d keep an eye on is the Stoopwriter name. It has a nice real-estate feel, but it also keeps the product framed as a writing tool. If this grows into a broader local SEO / visibility platform for agents, a cleaner SaaS-style name like Beryxa.com may age better.