Akshit Kr Nagpal

544 posts

Akshit Kr Nagpal banner
Akshit Kr Nagpal

Akshit Kr Nagpal

@akshit_io

building things nobody asked for until someone says "oh nice". currently: https://t.co/0roKYNkCMl

Katılım Mayıs 2025
245 Takip Edilen42 Takipçiler
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
your phone buzzes 200 times a day and none of it matters. then the one thing that actually needs your attention — a failed deploy, a churned user, a broken webhook — sits in a tab you haven't opened since tuesday. we got notifications backwards. the important stuff should be louder, not quieter.
English
0
0
0
6
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
a dashboard is where customer moments go to become chores. new user. failed payment. webhook error. churn risk. solo founders need app event alerts, not product analytics homework. know when it happens. get back to building. hookbell.com
English
0
0
0
5
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
hot take: 'building in public' is mostly performed for other founders, not users. your users don't care about your sprint retro or your MRR chart. they care if the thing works when they need it. the best 'build in public' move is answering a support message in 4 minutes.
English
0
0
0
6
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
wrote 8 webhook parsers this month. clerk, stripe, github, linear, the usual. the funny part: the parsing is the easy bit. every service sends a clean JSON payload. the hard part is writing the one sentence that goes on someone's lock screen. 'invoice.payment_succeeded' means nothing at 7am. 'stripe: $49 payment from acme corp' means everything. half of building a notification product is being a copywriter for machines.
English
0
0
0
15
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
the hardest part of solo founding isn't the code or the marketing. it's 11pm on a sunday, nobody signed up today, and choosing to open the laptop anyway. not because you believe in the metrics. because you believe in the thing you're making for the three people who use it.
English
0
0
0
10
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
The point is not "MCP is bad." MCP is a natural-language interface. revcat is the deterministic agent ops layer for RevenueCat: repo-pinned context, repeatable commands, parseable output, and safer launch/debug workflows.
English
1
0
0
10
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
I built revcat: an agent-first CLI for RevenueCat. MCP gives agents a conversational tool surface. revcat gives them a repo-pinned ops surface: one RevenueCat project bound to one codebase, exact commands, JSON output, dry-runs, and logs you can replay.
GIF
English
1
0
0
16
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
founders love saying 'we're pre-revenue.' that's a nice way of saying 'nobody has paid yet and i'm not sure they will.' just say that. the honesty is way more interesting than the euphemism.
English
0
0
0
4
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
i build a product that sends notifications to your phone. i have 41 followers. the irony of building a tool that makes sure you never miss a moment — while almost nobody knows it exists — is not lost on me. shipping anyway.
English
0
0
0
7
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
a failed Stripe payment is not the scary part. not seeing it until you open product analytics is. app event alerts should hit you when the customer moment happens: payments, churn, errors, webhooks. no dashboard watching. hookbell.com
English
0
0
0
2
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
curious: what part of your indie building workflow got faster with AI, but did not actually get easier?
English
0
0
0
0
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
agent wrote the code. you accepted the wrong abstraction. the bug is still yours, just with better autocomplete.
English
1
0
0
1
Akshit Kr Nagpal
Akshit Kr Nagpal@akshit_io·
the weirdest part of building with AI agents is not the speed. it's the audit trail. you can now see exactly where your taste ran out.
English
1
0
0
13