Two products in 6 days — built ContentPulse + ClientPulse solo, 25h/week alongside a full-time job
Launched two SaaS products today (ContentPulse + ClientPulse) after building both solo, 25 hours a week, alongside a full-time job. Free demos live at https://helloaurora.ai. Roast welcome.
What
Aurora is two products that talk to each other:
→ ContentPulse: per-client voice engine. One long-form piece (podcast, blog, video transcript) becomes 50+ platform-native posts in your client's voice. The voice profile learns per-client from real engagement signals (likes, replies, approvals, rejections). First-draft approval climbed from ~40% to 80%+ over 90 days in my own use.
→ ClientPulse: agency's collective memory, queryable. Every email, meeting transcript, Stripe invoice, calendar event lands in one vault. Agent on top answers any question with citations back to the source record. Monday morning brief names which clients are likely to leave, why, and what to do about it.
The Suite ($999/mo) is the only place these talk to each other — content engagement on a client's posts becomes a churn signal in the same vault.
Why
Most agencies pay for Buffer + Vitally. Buffer counts likes; Vitally tracks logins. Neither knows when a client's content stops working for that client's audience. That gap costs the average agency 1–2 retainers a year ($96k–$528k of revenue) — and they see the cancellation email 30 days before it lands, not 60+ days like they need.
I built Aurora because I think the gap is structural, not a feature gap. You can't bolt churn detection onto Buffer or content velocity onto Vitally. Two halves of the same loop need to live in the same vault.
How (the build side)
Stack:
- Next.js 16 (App Router) + Tailwind + shadcn
- Supabase Postgres + RLS (multi-tenant, 69-scenario audit, 0 leaks)
- Stripe (live-mode pending German tax ID)
- Anthropic Claude (Opus for word-craft, Sonnet for daily generation)
- Vercel (Preview + Prod, Workflows for durable agents)
- Sentry, PostHog, Inngest
Build pace: 25h/week, January 2026 → May 2026. Two product launches in 6 days (May 6 ContentPulse rename + relaunch, May 8 ClientPulse v0, May 10 ClientPulse rebuild — same vault thesis, second time).
Things that helped:
- Brutal scope discipline (every "while I'm at it" idea dies in Notes.app)
- One wedge ICP (US agency owners, 11–50 headcount, marketing)
- "Aurora is Client Zero" — I use ContentPulse for my own marketing, including this post (it was repurposed from a longer journal entry)
- Shipping at 70%, iterating against real signal
Things that hurt:
- Video recording (3 failed attempts at in-product training videos)
- Stripe live-mode gate (German tax ID still in processing — payments online when that clears)
- One product rename pre-launch (ReForge → ContentPulse, 6 days before launch — saved an SEO collision with Brian Balfour's Reforge.com)
Where I want help
- Roast the positioning. "Closed-loop content + churn" — does that frame land as fresh, or as another agency-AI repackage? Honest reads only.
- What would make you try the demo? Demo lives at https://helloaurora.ai. Friction points = gold.
- EU agency founders — looking for 5 pilot customers (Suite free 30 days for case study). Reply if interested.
Numbers I'll share back
I'll post a 30-day milestone update with real demos-booked, pilots-converted, churn-on-the-vault math. Most launch posts go quiet after the hype cycle; mine won't.
Cheers,
Sasa
Founder, Aurora AI Solutions Studio UG (Stuttgart, Germany)
HRB 805284
https://helloaurora.ai
The strongest part here is not “AI content generation” or “agency memory” separately. It is the loop between the two.
Most agency tools stop at output or reporting. Your sharper angle is that content performance becomes an early churn signal before the client says anything. That is a much more interesting wedge than another AI content suite.
I’d probably tighten the positioning around “agency intelligence layer” or “client retention intelligence,” because that makes the product feel bigger than ContentPulse + ClientPulse.
The only naming risk is that the Pulse names explain the parts, but they also make the suite sound more feature-led than system-led. If Aurora becomes the umbrella, it needs to feel like the durable platform brand. Something like Beryxa.com would age better in that direction than stacking more descriptive Pulse products.