screenpipe
the context layer for AI agents
your screen shows what you're doing. AI agents need to know what you're doing.
we capture the screen, infer intent, and let agents work for you in the background.
the insight
today: prompts are a tax
- → you stop working to prompt AI
- → you translate intent to text (lossy)
- → you wait for response
- → one prompt at a time = bottleneck
tomorrow: screen removes the tax
- → AI watches what you do
- → intent inferred from screen (lossless)
- → agents spawn in parallel
- → you review outputs, not write inputs
screen is the highest-fidelity signal of human intent. it shows what you do, not what you remember to type.
market timing
rewind
$350M raised
acquired by meta. shut down.
limitless
VC-backed
acquired by meta. shut down.
screenpipe
open source, local-first
only VC-backed company left.
meta spent $350M+ acquiring both competitors, validating the market. microsoft recall = windows-only, enterprise, privacy nightmare.
traction (feb 2026)
real DAU
60-69
2x in 2 weeks
revenue
$4.7K
/week (incl. lifetime)
activation
31%
was 19% two weeks ago
real WAU
286
intentional users
github stars
16.6K
CAC
$0
organic only
runway
32+
months
power user
604
shortcuts/week
the honest problem
31% activation — was 19% two weeks ago
461 people open the app weekly. 142 do something intentional. shipped shortcut-habit onboarding → DAU doubled (35 → 69) in two weeks. top user does 604 shortcuts/week. when it clicks, it's addictive.
product works. activation is moving fast. one onboarding change doubled DAU.
what's broken
- • 69% of openers still don't engage
- • permission loss (93 users/week)
- • windows perf (41% CPU)
what I'm doing
- • shortcut onboarding (shipped, working)
- • 3+ user calls/week
- • stability fixes daily
the trajectory
- • activation 19% → 31% in 2 weeks
- • DAU 35 → 69 in 2 weeks
- • target: 40% activation
used by people at
Thomson Reuters
$60B market cap
Planet Labs
NYSE: PL
NVIDIA
NASDAQ: NVDA
Broad Institute
MIT / Harvard
Cafe24
KRX: 042000
GMO Internet
TYO: 9449
Google
Chromium team
+ universities, startups, indie devs
go-to-market
what we do
- → 100% organic distribution
- → github, reddit, twitter, HN
- → daily video content
- → fix activation (main lever)
- → agent integrations (new audience)
what we don't do (yet)
- paid ads (don't know unit economics)
- sales team (no enterprise PMF)
- B2B outbound (retention not proven)
- partnerships (distraction)
roadmap (grounded)
Q1 2026
fix activation + stability
Shortcut onboarding shipped → DAU 2x. Now fixing permissions, Windows perf, retention.
$30-50K MRR
Q2 2026
first hire
Senior engineer. Scale shipping velocity.
$50-75K MRR
Q3-Q4 2026
team of 3
Rust engineer + designer. Retention proven.
$100K+ MRR
2027
series A
At $1M ARR. Scale team, enterprise exploration.
$1M ARR
the vision
the vision: you work → AI swarm works for you
waiting
your screen
Zoom - Sales Call with Acme Corp
Y
S
Live
“...excited to move forward. Let's schedule the demo for next Tuesday...”
screenpipe
spawning agents
ai workforce (parallel)
email
idle
notes
idle
calendar
idle
crm
idle
no prompts needed|parallel execution|you review, AI executes|you are the CEO
founder
louis
solo founder · @louis030195
why I win
- → I ship: 51 commits this week, 12 releases
- → I don't quit: survived leukemia at 13
- → I'm building what I want: 20K notes, 100+ agents
- → I sell without selling: $4.7K/week revenue, $0 CAC
time allocation
40% coding (activation)
25% user research
20% distribution
15% ops + hiring
14K+ contributions last year
the ask
not raising now. fixing activation — and it's working. DAU doubled in two weeks from one change.