SAASPOCALYPSEverdict #SOCIALFETCH-F7AA
scanned 2026.04.28 · 14:18
subject of investigation
socialfetch.dev
▸ unified social media scraper API
verdict: MONTH
buildability score
35
/100
tier · month
the blunt take
“You're not building a social API — you're building a maintenance contract with 20+ platforms that change their DOM every Tuesday. The product is the uptime, not the JSON.”
The engineering is real: residential proxies, per-platform parsers, transcript extraction, anti-bot evasion, and a normalized envelope that survives Instagram's next redesign. That's months of work, not a weekend — and it never stops being work.
cost breakdown.
their price ←→ your price
what they charge●
Starter pack (one-time)
$25
/ one-time / 10k credits
※ Pay-as-you-go credit packs; no subscription. Growth pack $99 for 50k credits.
annual:usage-based
what it costs you✦
01 · Vercel Pro (API gateway + edge)$20.00
02 · Supabase free (API keys, usage tracking)$0.00
03 · Residential proxy network (BrightData / Oxylabs)??? — scales with volume
04 · Cloudflare R2 (response caching, media)$1.00
05 · Resend (alerts, onboarding emails)$0.00
06 · Domain$1.00
07 · Sentry free (scraper error tracking)$0.00
08 · Railway / Render (scraper workers)$7.00
09 · Playwright / Puppeteer infra (browser pool)$24.00
TOTAL / mo$53.00 + usage
▸ break-even:immediately at any usage — $25 one-time buys 10k requests, and your infra costs start at ~$50/mo before you've scraped a single profile reliably
or, you know, use one of these.
if building feels spicy
option A
Apify (self-hosted actors)
Open-source scraper actors for TikTok, Instagram, YouTube. Run on their infra or yours. Massive head start.
option B
RapidAPI Marketplace (resell)
Aggregate existing social APIs under your own key. Less engineering, more margin math.
option C
Bright Data + custom parsers
Buy the proxy layer, write the parsers yourself. Cheaper than building proxy infra from scratch.
what'll actually be hard.
est. total: 3 months
▸ 2 weeks on auth + unified schema · 6 weeks building 5–6 platform scrapers · 4 weeks on proxy infra + anti-bot · ongoing: patching broken selectors forever
easy
medium
hard
nightmare
01
easy
Unified JSON envelope
Design a normalized schema once. Boring but doable in a day.
02
easy
API key auth + credit metering
Supabase + a middleware counter. Standard stuff.
03
medium
Per-platform parsers (YouTube, Reddit, Bluesky)
These have public APIs or lenient rate limits. Start here.
04
hard
Anti-bot evasion for Instagram & TikTok
Fingerprinting, rotating user agents, residential IPs, CAPTCHA solving. This is the real product.
05
hard
Video transcript extraction
YouTube has captions API. TikTok does not. You're running Whisper or paying for it.
06
nightmare
99.8% uptime across 20+ platforms
Every platform redesign breaks your parsers. This is a full-time on-call rotation, not a side project.
recommended stack
Next.js (API routes + dashboard)Playwright + Puppeteer (browser pool for JS-heavy platforms)Supabase free (API keys, billing state)Bright Data / Oxylabs (residential proxies)Railway workers (scraper queue)
ready to build?
We'll email you the build guide. Cancel some plans.
▸ generated with love, by a heartless robotverdict v2.1 · saaspocalypse.dev