halcy

the OS for social travel

02. what we're building

the OS for social travel.

halcy is a social-first agentic platform for travel, built to help people discover, plan, experience, and remember real-life adventures together. one place for the entire journey: from the first spark of inspiration shared between friends, to the living memory that seeds the next trip.

"travel is inherently social, but every tool treats it as a solo transaction. halcy starts with people: tribes of your existing friends, and collectives of like-minded travellers. sharing memories, inspiring each other, and turn ideas into action from one-click-buy travel plans. the social layer is the product."

03. why

two forces converging.

group travel is fundamentally broken

group travel is chaos. whatsapp threads, google docs, airbnb, tripadvisor, skyscanner, splitwise; none talk to each other. the cognitive tax on connection is so high that plans collapse before they begin, or even worse, travelers settle for suboptimal plans.

277

web pages visited before booking a single trip (expedia)

89%

of travelers frustrated when planning online (amadeus)

google docs

airbnb

tripadvisor

discretionary time will explode

as AI-led productivity gains penetrate white-collar work, discretionary time will expand. AI users already save one workday per week (LSE), and adoption is accelerating fast.

gen z: loneliest generation

AI surpasses humans in 50% of tasks

white-collar displacement

the most important, and final, unsolved problem becomes: how do people spend that time meaningfully?

04. why now

four drivers enabling this moment.

legacy penalty and channel conflict

airbnb, booking, tripadvisor are penalised by decade-old architectures. their business models create structural channel conflict. airbnb can never surface booking.com inventory; booking can never recommend an airbnb. a holistic solution is impossible by design.

agentic AI enables the product, now

agentic AI models that autonomously execute multi-step workflows across APIs only became viable in the last 12 months.

social media's structural failure

the incumbent for discretionary time is social media: inspiring in small doses, catastrophic at scale. it creates the loneliness it claims to cure. halcy uses the positive signals of inspiration and connection, but doesn't amplify doom scrolling.

the experience economy

secular preference shift from materialism to experiences. 78% of millennials choose experiences over things (WEF) and experiential travel is growing at 10.5% CAGR.

05. the opportunity

a €10.4tn+ market with no single true owner.

~€10.4tn TAM

total travel market in 2025, already 10% of global GDP (WTTC / Oxford Economics)

the incumbents

airbnb, hotels.com, booking are vertically focused. channel conflict structurally prevents full-world inventory or holistic solutions.

momondo, expedia, large scale aggregators with no curation. the research still needs to be done by the traveller. each traveller sees a different price point at different points in time.

trip.com and tripadvisor attempted the holistic angle and both failed; tripadvisor became a review farm, trip.com never cracked the western consumer.

all lack two critical flywheels: (1) no social features; (2) limited quality real-world signal from feedback loops. fundamentally transactional databases.

the AI-first travel challengers, and large scale LLMs

mindtrip ($22.5M), airial ($3M), layla ($3.3M), have reached +$29M funding combined proving the market is hungry. but all miss the mark.

all are chat-window interfaces: no mobile UX, no collaboration, no community, no memory loop. fundamentally wrappers with no proprietary data moat.

large-scale LLMs (claude, chatGPT, gemini) are antisocial by design: only targeting the single-user. google tried social with google+ and failed. google canvas limited by single-user architecture.

06. the solution

one platform. every step of the journey.

THE JOURNEY

01 · idea

discover destinations via social feed of real trip stories

02 · coordinate

AI resolves budget, dates & preferences across your group

03 · book

one-click group booking, full inventory, splits built in

04 · remember

shared photo vault, AI recaps, live ratings feed the reality graph

05 · inspire

memories inspire others — communities form — flywheel restarts

07. technology

the reality graph.

global world inventory from day 1. social data flywheel compounding over time from trip ratings, shared images and memories, tribe behaviour, and group preferences.

live integrations

gemini duffel viator ticketmaster places foursquare mapbox unsplash youtube open-meteo

pending, code ready

stripe booking.com expedia getyourguide klook kiwi seatgeek eventbrite opentable thefork trainline rome2rio uber cartrawler

architecture and data advantages

agentic execution engine

a tool-calling agent with 16+ specialised tools autonomously queries live supplier APIs (flights, stays, activities, events, restaurants), resolves group logistics, and executes booking. AI generates and optimises itineraries with full trip context, and the reality graph, not a chatbot bolted onto search.

group-first data model

trips, itineraries, and member travel are modelled as shared, multi-actor objects. each traveller has independent flights, dates, and preferences within a single trip. group decisions (which stay, which activities) are resolved through democratic Borda-count voting with automatic tally and apply.

social data flywheel → reality graph

tribes and communities generate continuous signal with trip completions and live ratings creating high quality data. this feeds a confidence-scored place knowledge base: anti-gaming by design. 36k+ curated places already seeded. every completed trip and community mention compounds the signal.

personalised feed and discovery

a multi-stage scoring pipeline ranks content by social graph proximity, including: destination relevance, engagement, author credibility, and freshness, with an exploration budget for serendipity.

08. meet hal

meet hal.

books on your behalf

hal is your AI travel companion, an agentic intelligence that doesn't just answer questions, but acts on your behalf. hal orchestrates the entire journey: resolving group preferences, searching inventory, booking, and adapting plans in real time.

🧠

reasons

resolves conflicting preferences across group members using multi-step reasoning

⚡

agentic

autonomously queries APIs, compares options, and executes bookings. not just suggestions

🔄

learns

every trip, rating, and shared memory refines hal's understanding of your tribe

agentic, not assistive. it proactively coordinates schedules, flags conflicts, suggests alternatives, and drives conversion to booked trips. powered by the reality graph, hal gets smarter with every interaction.

09. the app

the app.

📱

social feed

🫦

booking flow

🤖

hal

10. go-to-market

how we see this playing out.

PHASE 01

now

Continue building out the reality graph, collaborative, agentic and generative capabilities. Target feature superiority vs all incumbents across critical features in a multi-person environment before official launch.

PHASE 02

launch

Create initial critical mass through viral social media marketing to drive grassroots early adopter base. Snowball effect of people inviting friends into trips takes it from there.

→ why users join

Truly collaborative, lowers barriers to plan experiences with friends. Chaos mode specifically designed as a virality enabler.

→ why users stay

Agentic & generative capabilities, deep personalisation, embedded memories, and helpful communities.

PHASE 03

compounding

network effects

User base scales. Halcy becomes a real funnel for aggregators. Those without API integrations are forced to approve, losing channel-share to competitors who've deeply integrated. User, trip and community data power the reality graph over time. Each new user joining gets a better value proposition than the last.

PHASE 04

end-state

value capture

Aggregators become background plumbing. Trips are planned, booked and completed end-to-end in Halcy. Pricing power vs aggregators increases as Halcy sits in the cockpit of travel decisions. At sufficient scale, Halcy can collapse the value chain and own the aggregation layer itself.

each phase compounds into the next → the moat deepens with every user, every trip, every memory.

11. team

built by obsessed people.

LG

leo gustafsson

co-founder

MSc engineering student (KTH)

AÅ

alva åqvist

co-founder

5 years in tech investing at EQT (2021–26), ex-mckinsey (2019–21). MSc in applied mathematics (KTH)

🧢

co-founder

5 years in tech investing and consulting. BSc stockholm school of economics

get in touch

alva.aqvist@halcy.site leo.gustafsson@halcy.site