SAASPOCALYPSEverdict #LOOM-55CD
scanned 2026.05.04 · 14:34
subject of investigation

loom.com

async screen recording & video messaging
verdictCONTESTED
wedge score
61
/100
wedge thesis

the door is niche distribution: Loom now lives inside Atlassian's enterprise motion, leaving the solo-dev, creator, and SMB segments underserved by a product that's drifting upmarket.

real walls — pick your flank·ship in 6 weeks·run for $42.00 + usage
the doorregulatory
wedge

where the walls are.

methodology →
the door

no regulatory wall — SOC 2 doesn't count.

watch out

the technical wall is real — research-grade engineering, not a weekend.

capital
4.0/10
investment the incumbent had to make
why this scoremedium confidenceLoom's core product is software-only with no meaningful physical infrastructure, inventory, or payments risk. The...

Loom's core product is software-only with no meaningful physical infrastructure, inventory, or payments risk. The Atlassian acquisition adds enterprise sales motion and compliance overhead (SOC 2, enterprise procurement), but the underlying infra (CDN, object storage, video encoding) is all commodity cloud. The desktop app requires a kernel extension / virtual audio driver on Mac, which is a real engineering cost but not a capital moat per se. An indie builder can replicate the infra stack for ~$42/mo as the report itself demonstrates.

  • Estimated competing infra cost is $42/mo + usage — entirely commodity cloud (Vercel, R2, Mux, Supabase)
  • Loom's own stack is Vercel + CloudFront — no proprietary infra
  • Atlassian acquisition adds enterprise compliance overhead but no capital-intensive physical assets
technical
5.0/10
depth of the underlying engineering
why this scoremedium confidenceThe core recording loop is explicitly described as a weekend build using standard browser APIs. The real technical...

The core recording loop is explicitly described as a weekend build using standard browser APIs. The real technical depth lives in two places: (1) the Mac desktop app requiring a kernel extension / virtual audio driver for system audio capture, and (2) the AI-powered Jira ticket feature that injects into pages to capture console errors and network activity — a genuine browser extension security boundary problem. These are non-trivial but not fortress-level; they are well-scoped engineering problems a small team can solve in weeks to months, not years.

  • getDisplayMedia() + getUserMedia() screen/cam capture is documented as 'easy' — MDN has the full example
  • Chrome Extension MV3 rated 'medium' — annoying but well-documented, ~1 week
  • Desktop system audio capture rated 'hard' — requires kernel extension / virtual audio driver on Mac, real pain point
network
3.0/10
users compound users
why this scoremedium confidenceLoom has a mild viral loop: every shared video link is a passive ad, and recipients are nudged to sign up. However,...

Loom has a mild viral loop: every shared video link is a passive ad, and recipients are nudged to sign up. However, there is no true marketplace, no social graph, no UGC flywheel, and no multi-sided liquidity. The network effect is weak one-sided virality (sender → viewer → potential signup), not a defensible network. Atlassian's ecosystem integration adds some partner/app network value, but that accrues to Atlassian, not to a standalone Loom moat.

  • Share-a-link model creates passive viral distribution but no locked-in social graph
  • No marketplace or multi-sided platform identified
  • Atlassian integration adds ecosystem leverage but is an acquirer asset, not a standalone Loom network moat
switching
4.0/10
stickiness of customer data + workflow
why this scoremedium confidenceSwitching cost is moderate. Users accumulate a library of recorded videos, links, and comments that live on Loom's...

Switching cost is moderate. Users accumulate a library of recorded videos, links, and comments that live on Loom's platform — migrating that archive is painful. Workflow lock-in exists for teams that have embedded Loom links in Notion docs, Jira tickets, Slack threads, etc. However, the report's own wedge thesis is that Loom's upmarket drift is leaving SMB/indie users behind, implying those users haven't yet built deep lock-in. Enterprise users with Jira integration have higher switching cost, but that's the segment being ceded.

  • Accumulated video library and shared links create migration friction
  • Loom links embedded in Jira, Notion, Slack create workflow lock-in for existing users
  • Wedge thesis targets solo-dev, creator, SMB segments — these users have shallower lock-in than enterprise
data
4.0/10
proprietary data accumulates over time
why this scorelow confidenceLoom has accumulated a large corpus of screen recordings, transcripts, and behavioral engagement data (watch time,...

Loom has accumulated a large corpus of screen recordings, transcripts, and behavioral engagement data (watch time, replay points, comment anchors). This could theoretically train better AI summarization or anomaly detection models. However, there is no public evidence that Loom has built proprietary models trained on this corpus, and the AI features described (Whisper transcription, Jira ticket generation) appear to use commodity LLM APIs. The data exists but there is no evidence it has been weaponized into a defensible flywheel.

  • Large corpus of screen recordings and transcripts exists by virtue of scale
  • Timestamp-anchored engagement data (watch heatmaps, replay points) is behaviorally rich
  • AI transcription uses OpenAI Whisper API — commodity, not proprietary model
regulatorydoor
2.0/10
real licenses, not SOC 2 theater
why this scoremedium confidenceLoom is not a regulated product. There are no financial licenses, no clinical/EHR data obligations, no KYC/AML...

Loom is not a regulated product. There are no financial licenses, no clinical/EHR data obligations, no KYC/AML requirements, and no money transmission. SOC 2 compliance is mentioned implicitly via enterprise context but SOC 2 alone is explicitly low per the rubric. The browser extension security boundary (console/network capture) touches enterprise security policy but is not a regulatory moat — it's a trust and procurement friction, not a license.

  • No financial, healthcare, or identity-regulated workflows identified
  • No KYC/AML, HIPAA, FINRA, or PCI obligations described
  • SOC 2 likely required for enterprise sales but explicitly excluded from high scores per rubric
distribution
4.5/10
brand SERP grip, knowledge graph, news flow
take

the blunt take.

Loom got acquired by Atlassian and is visibly pivoting toward Jira-integrated enterprise workflows. That's a gift: the free-tier indie user and the small async team are now second-class citizens on a product roadmap written for Fortune 500 IT buyers.

The core recording loop — capture screen + cam, auto-upload, share a link — is a weekend build. The moat was always distribution and habit, not technology. With Loom's center of gravity shifting to enterprise, the habit layer is up for grabs again.

cost

cost of competing.

what they charge
Business plan
$12.5
/ user/mo
Starter is free with 5-min limit; Business ~$12.50/user/mo billed annually
annual:$150
what running yours costs
01 · Vercel Pro (video landing pages + API routes)$20.00
02 · Cloudflare R2 (video storage, light usage)$1.00
03 · Cloudflare Stream or Mux (video encoding/delivery)$20.00
04 · Supabase free (auth + metadata DB)$0.00
05 · Resend free (share notifications)$0.00
06 · OpenAI Whisper API (transcription)??? — scales with recording minutes
07 · Domain$1.00
08 · Sentry free (error tracking)$0.00
09 · OAuth (Google, GitHub)$0.00
10 · Chrome Web Store developer fee (one-time, amortized)$0.00
TOTAL / mo$42.00 + usage
▸ break-even:immediately at the paid tier — Loom's Business plan runs ~$12.50/user/mo; your infra runs ~$53/mo flat regardless of seat count, so any team of 5+ is cheaper on day one.
build

what you're up against.

1 week Chrome extension + desktop capture · 1 week upload pipeline + CDN · 2 weeks player, sharing, comments · 1 week AI transcript · 1 week polish + auth
easy
medium
hard
nightmare
01
easy
Screen + camera capture in browser
getDisplayMedia() + getUserMedia(). MDN has the full example. Works in Chrome day one.
02
easy
Upload pipeline to R2/S3
Presigned URL upload from the client. Supabase storage or R2 both have SDKs. One afternoon.
03
medium
Chrome extension with one-click record
Manifest V3, background service worker, content script injection. Annoying but well-documented. A week.
04
medium
Video player with comments + emoji reactions
Timestamp-anchored comments require a custom player (Video.js or Plyr + overlay layer). Not hard, just fiddly.
05
hard
Desktop app (Electron/Tauri) for system audio capture
Browser APIs can't capture system audio on Mac without a virtual audio driver. Loom ships a kernel extension. This is the real pain point.
06
nightmare
AI-powered Jira ticket generation with console/network capture
Loom's new flagship feature injects into the page to capture console errors and network activity alongside the recording. This is a browser extension security boundary problem — non-trivial and the reason enterprises pay.
stack

their position.

detected signals· measured
hostingVercelframeworkNext.jscdnCloudFront
recommended stack · inferred
inferNext.js 15 (marketing + viewer pages)inferCloudflare R2 + Mux (storage + adaptive streaming)inferSupabase (auth + Postgres metadata)inferElectron or Tauri (desktop app for system audio)inferOpenAI Whisper (transcription) + Chrome Extension MV3
rivals

who else has tried this.

option A
OBS + Streamable/YouTube unlisted
Free, zero infra. Overkill for recording but works for async sharing if you don't need a viewer with comments.
option B
Cap (open-source Loom clone, self-host)
Already exists on GitHub. Electron + S3. Fork it, brand it, ship it faster than building from scratch.
option C
Tella.tv or Screenpal free tier
If you just need to skip the build — both are cheaper than Loom Business for small teams.
compare

similar scans.

same shape - different moat
ready to wedge in?
Get the wedge plan. Cancel some plans.
▸ generated with love, by a heartless robotverdict v2.5 · saaspocalypse.dev