I'm Rasif Ali Khan. I run GFX Bracket in Pakistan. I'm a product owner who rebuilt File Transcribe from WordPress, stubbornness and one too many plugin conflicts included.
This post is part of a series with one spine: how each tool in my stack helped me reach 100 paid subscribers after escaping WordPress. Vercel is where that story starts, because if you can't deploy safely, you can't grow.
The headline: 100 paid subscribers on the v2 stack
File Transcribe now has 100 paid subscribers.
That's not a forecast for July 2026. That's where the rebuild landed, people paying monthly for transcription on Next.js + Vercel + Supabase + Paddle, after v1 on WordPress proved the idea but couldn't carry the weight.
The first version made $87.57 through Paddle. Tiny number. Massive signal: strangers would pay. The modern stack turned that signal into a hundred paying customers, and the infrastructure to add more without FTP panic.
"The product was not a failure because the idea was bad. The architecture was too fragile to become a professional SaaS."
Vercel didn't magically create subscribers. It removed the ceiling, deploy in minutes, preview before merge, rollback when I break billing, so I could spend my weeks on product and growth instead of nginx.
Full timeline (WordPress → Vercel → 100 subs)
| When | What happened |
|---|---|
| v1 build | WordPress, custom PHP plugins, WPCode, ACF, server-side transcription, MyCred minutes, Paddle experiments |
| First revenue | ~$87.57 through Paddle, proof strangers pay |
| AppSumo | Rejected, forced quality and positioning rethink |
| Product Hunt v1 | Launched on WordPress, traffic, learning, stress |
| David Mbugua (YouTube) | 3 videos on WordPress v1, still motivates v2 polish · channel |
| Breaking point | Plugin conflicts, key exposure risk, job timeouts, CSS patching instead of features |
| GFX dashboard | Internal app on GitHub + Vercel + Supabase, first professional deploy loop |
| v2 rebuild | Clean repo, Vercel previews, serverless APIs, Paddle webhooks |
| Growth on v2 | Homepage upload, guest `/try/`, library, Pro tiers, 100 paid subscribers |
| July 2026 | Product Hunt v2, public relaunch moment after proving paid growth |
WordPress hosting vs Vercel (why I couldn't scale to 100 on v1)
Deploy
- WordPress v1
- FTP, cache plugins, pray
- Vercel v2
- Git push → preview URL → merge → live
Preview
- WordPress v1
- Basically none
- Vercel v2
- Every branch shareable, I test billing before prod
API routes
- WordPress v1
- PHP + AJAX hacks
- Vercel v2
- Next.js Route Handlers + Serverless Functions
Background jobs
- WordPress v1
- Fragile HTTP
- Vercel v2
- Cron + queue, transcription survives tab close
Rollback
- WordPress v1
- Manual terror
- Vercel v2
- One click when Paddle webhook breaks
Global speed
- WordPress v1
- One bloated host
- Vercel v2
- Edge, US/UK uploaders matter
Secrets
- WordPress v1
- Theme/plugin risk
- Vercel v2
- Env vars only, never client bundle
Founder time
- WordPress v1
- Sysadmin
- Vercel v2
- Product + growth → 100 subs
On WordPress I was afraid to ship daily. On Vercel I ship multiple times a week. You cannot A/B pricing, fix checkout, or iterate onboarding when deploy day is a crisis.
What actually broke on WordPress (the unglamorous list)
Logic lived in:
- Custom PHP plugins (multiple versions in old folders)
- WPCode snippets
- Custom post type `transcription`
- ACF: file URL, transcript body, summary, translation, job ID, captions
- Sidebar "previous files" from WP_Query hacks
- Frontend JS → `admin-ajax.php` → transcription API
Every new feature touched five surfaces. MyCred minute credits fought membership plugins. Long jobs timed out. I couldn't swear users only see their own transcripts without unaudited PHP.
AppSumo rejection hurt. Competitors looked sharper. But the real blocker to 100 subscribers wasn't marketing, it was not trusting my own deploys.
GFX Bracket dashboard, the proof Vercel works for me
Before File Transcribe v2, I shipped an internal GFX dashboard: Google auth, Postgres, RLS, shared data, Vercel deploys.
That project taught me:
- Preview URLs before production
- Migrations in git, not phpMyAdmin
- One button deploy
When I created `file-transcribe` on GitHub, Vercel was decision #1. I'd already felt WordPress vs modern hosting. I wasn't guessing.
What runs on Vercel today (concrete)
| Surface | Role |
|---|---|
| filetranscribe.com | Marketing, blog, pricing, fast first paint |
| Homepage upload | Server routes, keys never in browser |
| `/auth/callback` | Supabase OAuth without managing Node servers |
| `/api/transcribe` | Server-side transcription routes |
| `/api/jobs/process` | Background transcription cron |
| Paddle webhooks | Verified signatures → `profiles.plan` |
| Blog series | These posts, outreach to Vercel, Supabase, etc. |
The mental shift: infrastructure invisible. I optimized guest UX, mobile playback, Ask AI presets, the things that convert free try → paid.
How Vercel specifically helped reach 100 paid subscribers
Fix checkout bug same day
- Without Vercel
- Fear
- With Vercel
- Deploy + rollback
Share preview with David / creators
- Without Vercel
- Impossible
- With Vercel
- Preview URL per branch
Handle PH traffic spike
- Without Vercel
- Plugin host meltdown
- With Vercel
- Edge + serverless scale
Ship blog + product same repo
- Without Vercel
- WordPress theme hell
- With Vercel
- One Next.js app
Iterate pricing copy
- Without Vercel
- Dangerous
- With Vercel
- Preview → merge
Subscriber #1 and subscriber #100 both needed reliable transcription. Subscriber #50 needed me not breaking auth during a plugin update. Vercel made that boring.
Results table
Deploy time
- v1 WordPress
- Days of friction
- v2 Vercel
- Minutes
Paid proof
- v1 WordPress
- $87.57
- v2 Vercel
- 100 paid subscribers
Post-deploy fear
- v1 WordPress
- High
- v2 Vercel
- Low
Time on infra vs product
- v1 WordPress
- 60/40 wrong way
- v2 Vercel
- 90/10 product
Personio wrote about leaving WordPress for Vercel and winning Core Web Vitals. I'm one founder in Pakistan, smaller scale, same relief: monolith out, product in.
For Vercel DevRel / marketing
Pitch summary:
- WordPress → Vercel migration (real pain, real tables)
- Solo founder, Pakistan, Paddle not Stripe
- Sensitive uploads, serverless + secrets discipline
- 100 paid subscribers on rebuilt stack, measurable outcome
- PH v2 July 2026, timely co-marketing hook
- Full series, Supabase, Next.js, Paddle posts link here
Republish-friendly. Happy to adapt quotes, screenshots, architecture diagram.
Try it
File Transcribe, upload audio or video, editable transcript, no signup required to try.
Related: Supabase → 100 subs · Next.js → 100 subs · Full stack hub
Rasif Ali Khan, Founder · GFX Bracket · File Transcribe