ai for the people who matter
warm.ai

networking
without the work.

warm is the ai that knows who you are, finds the humans that matter, introduces you, and keeps the relationship alive — while you're heads down.

live · your network
312
humans mapped
p1 · warm48
p2 · warm-ish76
p3+188
next warm touch
priya raman · tuesday 3pm
your network graph · 312 people · 48 warm
agent · running
raising a $3m seed
22 partners reached · 6 meetings · 2 term-sheets
warm hop path
to satya nadella
you
priya
james
satya
confidence: high · draft ready
priya raman · sequoia marcus chen · loop.ai ana silva · everstone james okafor · index lior mendel · modal sana khan · stripe yuki tanaka · deepmind amara okoye · everstone raj mehta · flux hana park · linear diego alvarez · quiet capital elena voss · northstar priya raman · sequoia marcus chen · loop.ai ana silva · everstone james okafor · index lior mendel · modal sana khan · stripe yuki tanaka · deepmind amara okoye · everstone raj mehta · flux hana park · linear diego alvarez · quiet capital elena voss · northstar

the problem

§ 01

profiles are static resumes.
humans are living context.

we still scroll linkedin. we ctrl-f email threads. we forget the person we met at the dinner in march. meanwhile, ai already understands humans better than any profile ever will.

there is no trusted ai with your full context — that speaks for you, finds the right room, and stays warm on your behalf. warm is that ai.

the three moves

understand. match. maintain.

one product. three loops. compounding, quiet, always on.

01 · understand

a living model of you.

upload your resume, link your linkedin, take a 3-minute quiz. warm builds a private model of your skills, goals, and context — refreshed every time you send an email or take a meeting.

resumelinkedin3-min quizemail signalcalendar signal
02 · match

the right humans.

in your network, or one warm hop away. see the shortest, most likely path to any person on earth.

query >
gpu infra founder, seed, wants co-founder
→ 6 matches, 2 in-network, 4 via 2 hops
03 · maintain

nothing goes cold.

warm nudges follow-ups, drafts catch-ups, and watches the calendar. p1 to p5 rankings, automatic or manual.

ana silvaquarterly · voice note draftp2
yuki tanakareply overdue · 10dp2
diego alvarezcheck-in frip3
bonus · circles

circles form themselves.

went to stanford? there's a circle. worked at everstone? another. yc w23? in the batch. warm draws them, tracks overlaps, and lets you invent your own.

stanford24everstone12yc w2318founders31investors22ai research15+ new circle

the personal crm

connect it all. once.

warm's agent quietly reads across every place your relationships live. context follows the person, not the platform.

auto-dedupe across sources
p1 → p5 rank the system learns
circle overlap map
your vault. zero-copy. no training on you.
linkedin
312 connections
email
gmail · outlook
calendar
90d meetings
whatsapp
38 threads
contacts
1,204 imported
more soon
signal · x · tg

agents that do the work

set a goal. walk away.

long-running agents keep pulling threads for you — cold and warm outreach, hop-by-hop introductions, calendar coordination. you review, they act.

goal
running
raise a $3m seed round
42%

22 partners · 6 meetings · 2 term-sheets

goal
running
find a technical co-founder
68%

shortlist of 8 · marcus chen frontrunner

goal
planning 3 hops
warm intro to satya at msft
12%

you → priya → james → satya

goal
paused for review
recruit 2 founding engineers
55%

5 candidates · call on lior?

goal
drafting
revive 12 cold p2 relationships
30%

voice notes queued · you approve

reach

every person on earth is 3 warm hops away.

cold when you must. warm when you can. warm surfaces the exact chain of humans who can introduce you — and drafts the message for each hop.

hops ranked by reply-ratecontext per hopauto-draft introsgentle nudges
target
sn
satya nadella
chairman & ceo, microsoft
you
you
you are here
1
priya raman · sequoia
1st degree · 4-day median reply
2
james okafor · index
priya's peer · shared board
3
satya nadella · msft
james sits on 2 boards with him

for teams · b2b

the people graph of your whole company.

query by skill, not job title. get the answer from someone's agent first — and the human when only a human will do.

step 01

ask the org.

q: what did we decide about ranker v3 on march 14?
searches slack · notion · decisions · transcripts
step 02

who to go to.

sara nur · vp mlhigh
owns ranker v3 postmortem
kwame osei · vp platformmed
ran the mar-14 incident
step 03

then the human.

agent handles the answer. warm books the 15 min only when judgment is needed.

340
people mapped
1.2k
tribal knowledge items
12s
median query time
73%
meetings avoided
zero-copy · your data stays in your tenant
private beta · limited seats

your warmest intro
is one prompt away.

join the private beta. we're onboarding operators, founders, and investors first.