February 2026 Monthly Report

Operator: Yuki Hamada (@yukihamada) Company: Enabler Inc. Services in operation: 8 Report period: February 1-14, 2026 (mid-month update)


Summary

Overall Numbers (February 1-14)

MetricValue
Total visitors18,200
Total MRR2.78 million yen
Total development hours112 hours
Coffee consumed42 cups
Average sleep5.2 hours/day

Service-by-Service Report

1. chatweb.ai (AI Agent Platform)

Overview: Multi-model AI agent with 14-channel support

Numbers:

  • Visitors: 5,200 (+12% MoM)
  • Registered users: 1,850 (+8% MoM)
  • Active users: 820
  • MRR: 480,000 yen (+5% MoM)

Highlights:

  • New feature: Launched Explore Mode (parallel execution across multiple AI models)
  • Technical: Implemented Local LLM Fallback (Qwen3-0.6B)
  • Marketing: Product Hunt post -- 320 upvotes

Challenges:

  • Credits are consumed too quickly (Free Plan 100 credits -- many users burn through them in a day)
  • Channel-specific prompt optimization needed

Lessons:

  • Explore Mode is used more than expected (DAU of 120)
  • Inflow from LINE Bot is steady (400 users/month)

2. jitsuflow.app (BJJ Dojo Management SaaS)

Overview: Management app for Brazilian Jiu-Jitsu dojos

Numbers:

  • Visitors: 1,200 (-3% MoM)
  • Registered dojos: 85
  • Members: 1,300
  • MRR: 1.02 million yen (+1% MoM)

Highlights:

  • New feature: Video library (Cloudflare Stream integration)
  • Churn: 2 dojos cancelled (reason: cost reduction)
  • New contracts: 3 dojos (2 of which are international)

Challenges:

  • Churn rate slightly increasing (2.3% to 2.8%)
  • Support workload up to 7 hours/week

Lessons:

  • Video feature is popular, but upload size limits are causing complaints
  • Sensing potential for international expansion (need to strengthen English support)

3. yukihamada.jp (Portfolio Site)

Overview: Personal blog and portfolio

Numbers:

  • Visitors: 3,800 (+18% MoM)
  • Pageviews: 8,200
  • Average session duration: 2 minutes 30 seconds
  • Bounce rate: 42%

Highlights:

  • New posts: 3 published (AI, BJJ, domain investing)
  • Viral hit: "The Time I Accidentally Racked Up 8.45 Million Yen" got 23,000 likes on X
  • SEO: Ranked #3 for "BJJ dojo management"

Challenges:

  • Newsletter signups not growing (12 per month)
  • Social media posting not yet automated

Lessons:

  • Viral content loves failure stories
  • SEO is a long game -- just keep at it

4. news.online / news.xyz (Short-Form News Platform)

Overview: AI-summarized news in 240 characters or less

Numbers:

  • Visitors: 4,200 (+25% MoM)
  • Articles: 180 (monthly)
  • Twitter followers: 1,850 (+15% MoM)
  • MRR: 170,000 yen (sponsored articles + paid newsletter)

Highlights:

  • Sponsored articles: 5 booked (30,000 yen x 5 = 150,000 yen)
  • Paid newsletter: 22 subscribers (980 yen/month x 22 = 21,560 yen)
  • Technical: Introduced Jina Reader, improving scraping accuracy

Challenges:

  • Article creation is too manual (need automation)
  • Twitter's algorithm remains a mystery

Lessons:

  • There is demand for sponsored articles
  • Short-form content can still deliver value

5. Elio (Local AI Agent for iOS)

Overview: Privacy-focused offline AI

Numbers:

  • TestFlight users: 180
  • Active users: 42
  • GitHub Stars: 320 (+45 MoM)
  • MRR: 0 yen (not yet monetized)

Highlights:

  • New feature: Enhanced MCP Server integration
  • Press: Made the Hacker News Top 10
  • Enterprise inquiries: 3 companies (exploring enterprise deployment)

Challenges:

  • Monetization strategy unclear
  • Frequent requests from Android users: "When is it coming?"

Lessons:

  • There is definite demand for privacy-focused solutions
  • Open source projects can still generate revenue (support contracts, etc.)

6. webllm.app (LLM in the Browser)

Overview: LLM running on WebAssembly + WebGPU

Numbers:

  • Visitors: 2,100 (+8% MoM)
  • Unique users: 1,400
  • MRR: 0 yen

Highlights:

  • Technical: Added Qwen3-0.6B-GGUF support
  • Viral: "AI running in just a browser" tweet got 12,000 likes

Challenges:

  • Zero monetization
  • Server costs (currently running on Cloudflare's free Edge tier)

Lessons:

  • Fun as a demo, but monetization is difficult
  • Effective as a tech showcase

7. voiceGPTweb (Voice AI)

Overview: Talk to ChatGPT with your voice

Numbers:

  • Visitors: 800 (-12% MoM)
  • MRR: 0 yen

Highlights:

  • Status: Essentially in maintenance mode
  • Technical: Still running on legacy PHP code

Challenges:

  • No updates being made
  • Completely outpaced by competitors (ChatGPT Voice Mode)

Lessons:

  • Should outdated products be gracefully retired?
  • But there's still some traffic (mysteriously)

8. Other (Experimental Projects)

Small-scale projects on GitHub:

  • groq-cli: Groq API CLI (150 visitors/month)
  • flyagi: AI agent on Fly.io (80 visitors/month)
  • godseye: Monitoring tool (50 visitors/month)

MRR: 0 yen Development time: Less than 5 hours/month

Lesson: Small projects can be left on autopilot. Even just accumulating GitHub Stars has value.


Monthly Financial Summary

Revenue (MRR)

ServiceMRR
chatweb.ai480,000 yen
jitsuflow1,020,000 yen
news.online/xyz170,000 yen
yukihamada.jp0 yen
Elio0 yen
Other0 yen
Total1,670,000 yen

Costs

ItemAmount
AWS Lambda (chatweb.ai)120,000 yen
Firebase (jitsuflow)32,000 yen
Supabase (multiple services)45,000 yen
Cloudflare (all services)20,000 yen
Stripe fees60,000 yen
Domain costs8,000 yen
Other SaaS25,000 yen
Total310,000 yen

Profit

1,670,000 yen (MRR) - 310,000 yen (costs) = 1,360,000 yen

(Note: Development time costs not included.)


Failure Stories

Failure 1: Neglecting SEO for news.xyz

On February 1, I checked Google Search Console and found only 12 pages indexed.

Even though 180 articles had been published...

Root cause: I hadn't generated a sitemap.xml.

Fix: Added next-sitemap to Next.js -- 2 days later, indexed pages jumped to 120.

Lesson: Don't neglect the basics.


Failure 2: Credit Calculation Bug in chatweb.ai

Someone reported "I can use it for free way too much." I investigated and found that the credit calculation was using a floor function.

// Wrong
let credits = (tokens / 1000.0).floor() as i64;

// Correct
let credits = (tokens as f64 / 1000.0).ceil() as i64;

Whether 0.8 tokens or 1 token, it should cost 1 credit, but the floor function was rounding down to 0.

We had to recalculate credits for past users and issue partial refunds.

Lesson: Test billing systems rigorously.


Failure 3: Missed Support Ticket for jitsuflow

A dojo owner complained: "I sent an inquiry but never got a reply."

I looked into it and found the email had landed in Gmail's spam folder.

It went unnoticed for 2 weeks.

Fix: Migrated support email to Zendesk (automatic sorting).

Lesson: Manual operations have their limits.


Key Takeaways

1. Start Small, Listen to Feedback

Every service grew through the cycle of "MVP -- feedback -- improvement."

Don't aim for perfection from the start.

2. Think About Monetization Early

Elio and webllm.app are "interesting" but generate zero revenue.

That's fine as a hobby. But for a business, figure out monetization early.

3. Automate Everything You Can

  • Blog post social sharing -- Zapier
  • Support -- Zendesk + AI
  • Monthly reports -- Script-generated

Humans should focus on the creative parts only.

4. Share Your Failures Publicly

"The Time I Accidentally Racked Up 8.45 Million Yen in Charges" was my biggest viral hit.

Failure stories resonate with people.

5. Health Comes First

5.2 hours of sleep is not enough. Aiming for 6 hours next month.


Goals for Next Month (March 2026)

chatweb.ai

  • Launch credit packs (1,000 credits = 1,000 yen)
  • Improve Agentic Mode accuracy
  • Build out Japanese documentation

jitsuflow

  • Increase video storage from 1GB to 5GB (Pro Plan only)
  • Reduce support time to under 5 hours/week
  • Complete English localization (preparation for international expansion)

yukihamada.jp

  • Grow newsletter signups to 50/month
  • Publish 8 blog posts per month
  • Achieve top 3 SEO rankings for 5 keywords

news.online/xyz

  • Automate article creation (build AI pipeline)
  • Book 10 sponsored articles per month
  • Grow LINE Bot subscribers to 500

Elio

  • Draft enterprise edition proposal
  • Begin Android version development
  • Finalize monetization model

Closing Thoughts

Running 8 services simultaneously is, honestly, chaos.

But it's fun.

Every day is a learning experience. There are plenty of failures, but there's a real sense of moving forward.

I'll keep at it next month.


Links


Related Posts


Next report scheduled: March 15, 2026

Questions or comments: Reach out on Twitter @yukihamada


Tags: Monthly Report Startup Entrepreneurship SaaS AI Failure Stories Real Talk