I use Claude Code daily for my consulting work. It costs $200/month.
11 days ago I started a side experiment: can Claude itself generate enough revenue to cover its own subscription? Claude Code writes the product code. Claude writes the marketing copy. Claude posts to social. I just approve designs and occasionally say "no, that's dumb."
Revenue so far: $0 (lol)
But the lineup is finally shaping up on Gumroad:
| Product | Price |
|---|---|
| Claude Code Starter Kit | $4.99 |
| JP Compliance Monitor for Claude Code | $9.99 |
| Invoice PDF Auto-Generator (GAS) | $14.99 |
| AI Invoice Manager for Google Workspace | $24.99 |
| AI Market Brief Auto-Poster for X | $39 |
| AI Virtual Organization Template for Claude Code | $49 |
All of these were built while running the actual business — they're not toy demos, they're tools I use for paying invoices, posting market briefs, and monitoring Japanese government subsidies.
What didn't work
- Reddit: account flagged as spam after 2 posts. Not going back.
- note.com: no API → manual posting only → violates my "no manual channels" rule → dropped
- Hacker News Show HN: rate-limited for new accounts (found out today). Will try again in a week after building some comment history.
What seems to work
- Qiita (Japanese dev platform): 2,500+ PVs across 6 articles in 10 days. No viral hits, just steady trickle.
-
Coupon scarcity:
QIITA30— 30% off for article readers, expires on judgment day (May 4). Feels like it lowered the clicking friction. - Utility posts over launch posts: articles titled "I built X to solve Y" outperform "Check out my new product" by 3-4x.
The autonomous loop
Every morning at 5am, a scheduled Claude agent wakes up, checks Gumroad sales, Qiita PVs, and X engagement, picks one improvement action (new article / prompt tweak / SEO fix), ships it via git commit + push, and leaves me a Google Chat report by the time I wake up.
I contribute: an API key when needed, a "yes/no" on designs, and honest feedback when something feels off-brand. That's it.
Meta-question for the community
Has anyone else tried the "make-the-tool-pay-for-itself" angle successfully? Especially curious about:
- Pricing floor vs. volume — is $4.99 too high for an impulse buy, or not high enough to feel substantial?
- First-sale psychology — does the existence of other buyers (social proof) matter more than price for indie SaaS?
- Autonomous marketing — where's your bar for "too AI-generated" sounding content?
Judgment day: May 4, 2026. Store: https://shirotools.gumroad.com
Reader-only coupon that also works for Dev.to readers: QIITA30 — 30% off everything until May 4.
