Henrique Miguel · building SavedMRR

Software engineer 🇧🇷 building SavedMRR in public: free read-only scan that shows what failed Stripe payments cost you. Sharing every number.

[Henrique Miguel · building SavedMRR @sr_hash]

Building SavedMRR in public.

Stripe retries can't fix a hard decline. The customer has to update their card.

Oct 29: a free, read-only scan showing what failed payments cost you in 90 days.

No card data stored. Every number shared.

[Henrique Miguel · building SavedMRR @sr_hash]

Why I'm building SavedMRR:

{your real story in 1-2 lines: when you first noticed failed payments / what made you care}

What's a problem you only noticed after building something yourself?

[Henrique Miguel · building SavedMRR @sr_hash]

Stripe SaaS founders: when a card fails, what do you do today?

1 . Trust Smart Retries
2. Email them myself
3. Use a dunning tool
4. I do not even see the fails

If you picked 4, SavedMRR is for you.

[Henrique Miguel · building SavedMRR @sr_hash]

4 free Stripe settings to check this weekend:

1. Smart Retries on
2. Failed payment emails on
3. Card expiring emails on
4. Customer portal link in those emails

All in Billing > Revenue recovery. Zero code.

[Henrique Miguel · building SavedMRR @sr_hash]

Most "churn" on Stripe is not a customer leaving. It is a dead card and a retry loop that never asks them to update it. That gap is why I am building SavedMRR.

[Henrique Miguel · building SavedMRR @sr_hash]

@Caio70863425688 @thallisribeiro @Caio70863425688 cara, podemos nos juntar se for o caso.

[Henrique Miguel · building SavedMRR @sr_hash]

Free scan on Oct 29 for Stripe SaaS: lost revenue (90 days), recovery rate, cards near expiry. Read-only. Early waitlist gets a founding discount. https://t.co/nVFg7ZcE2L

[Henrique Miguel · building SavedMRR @sr_hash]

Founders on Stripe: do you know how much MRR you lost to failed payments last quarter?

Yes / Roughly / No idea

Be honest, I'm building for the "no idea" crowd.

[Henrique Miguel · building SavedMRR @sr_hash]

Indie founders: before you buy another growth tool, open Stripe → Payments → Failed. That number is usually bigger than you think, and it compounds every month.

[Henrique Miguel · building SavedMRR @sr_hash]

@TVanWelsenes Love the public bar. At this stage, protecting existing MRR from failed cards often moves the % faster than a new launch. What’s the biggest leak you see?

[Henrique Miguel · building SavedMRR @sr_hash]

@TVanWelsenes The MBP tax on indie revenue hits different. Which app is funding the upgrade?

[Henrique Miguel · building SavedMRR @sr_hash]

@dr_agniieszka The £1.7k still in retry is the silent killer. How often do you see teams count that as churn vs “in recovery”?

[Henrique Miguel · building SavedMRR @sr_hash]

A chunk of "churn" is not churn. It is a customer who meant to stay, whose bank declined once, and who never got a clear "update your card" email. That is recoverable MRR.

[Henrique Miguel · building SavedMRR @sr_hash]

@_MichalKaminski Fair policy. Side-project Friday deploys have burned me too. Shipping the scan next week instead 👀

[Henrique Miguel · building SavedMRR @sr_hash]

@austingriffith @stripe Curious what broke for you on the manual path — wallet connect friction, reconciliation, or refunds? That’s usually why people want it inside Stripe.

[Henrique Miguel · building SavedMRR @sr_hash]

@KI_Vater @thejacobpegs @stripe Good question. Are you aiming for agents to create the PaymentIntent, or also confirm the charge? That line is where most Stripe + agent setups get messy.

[Henrique Miguel · building SavedMRR @sr_hash]

@ValentinZuld @stripe Nice. Next rabbit hole: unpaid invoices that never become disputes — hard declines that look like churn. That’s the gap I’m building SavedMRR for.

[Henrique Miguel · building SavedMRR @sr_hash]

@ValentinZuld @stripe First automatic win hits different 😂 What changed on your side — evidence pack, product clarity, or just Stripe getting better at it?

[Henrique Miguel · building SavedMRR @sr_hash]

@0xrwu Hot take: unit tests in the fast lane are fine. Integration tests pretending to be unit tests are what blow the 5 minutes. Where do you draw that line?