IRL

IRL · Founder Preview

Turn online communities into offline magic.

IRL is a lightweight operating system for real‑world gatherings: discovery, coordination, and follow‑up for the people who actually show up.

Target users
Creators, community leads, city builders, power hosts.
Core promise
Less group chat chaos, more faces you actually know in the room.
AI angle
Assistants that understand relationships, not just RSVPs.

Built from lessons across products like Foursquare, Airbnb, Partiful and the latest wave of AI‑native tools. Designed for 2026, not 2012.

IRL Session

Weekend in New York · 18 confirmed

Live runbook

People graph

See who actually overlaps across your Discord, group chats, and city.

Before

Adaptive invite rails

Tight controls for who can invite whom, tuned for intimacy vs scale.

During

Memory layer

Auto‑organized recaps, photos, and intros so each event seeds the next.

After

This is a concept snapshot, not final UI. The goal is opinionated tooling that respects social nuance, not yet another mass‑blast event link.

Product pillars

Three legs of the IRL stool: discovery, coordination, and continuity. Each is built to play nicely with the tools communities already live in.

1 · Discovery that respects context

IRL treats “who should be in the room” as a first‑class problem, not an afterthought bolted onto a calendar invite.

  • • Soft circles and trust‑weighted invites instead of open links.
  • • Signals from existing communities (Discord, WhatsApp, Telegram, email).
  • • City‑level intent: “who’s hosting dinner this Thursday in Brooklyn?”

2 · Coordination without group chat chaos

Hosts get a clear runbook, guests get exactly what they need and nothing more.

  • • Lightweight flows for RSVPs, plus‑ones, and “maybe” handling.
  • • Playbooks for recurring formats like salons, walks, and work sessions.
  • • Real‑time state: “who’s actually here”, “who’s on the way”, “who’s late”.

3 · Continuity between moments

Every gathering is an input to the next one: who met, what clicked, what should happen again.

  • • Automatic, opt‑in shared recaps instead of scattered photo dumps.
  • • Memory graph of who overlapped where, across time and cities.
  • • Prompts for “next best” gathering when the energy is still high.

AI native, socially aware

The goal is not “AI that writes event descriptions”, but assistants that actually understand your social graph, constraints, and appetite for surface area.

  • • Copilot for hosts that proposes guest lists, formats, and timings per city.
  • • Semantic understanding of “vibe” and boundaries, not just availability.
  • • Privacy by default: the model sees patterns, not raw private messages.

This leans on the 2025‑era foundation model stack: small, fast, local‑friendly models for on‑device inference, plus hosted LLMs for heavier social reasoning.

System principles

IRL over MAUs

Optimized for repeat, high‑quality in‑person touchpoints, not raw usage time or vanity metrics.

Network‑aware, not network‑effect‑addicted

Useful for a single host and their graph on day one, but capable of city‑scale density.

Interoperable by default

Plays nicely with calendars, messaging apps, and existing communities instead of trying to replace them.

Near‑term roadmap

  1. Private host console for “IRL power users” in cities like New York, SF, LA, London.
  2. Simple flows for recurring formats (dinners, walks, salons) with opinionated defaults.
  3. Lightweight guest experience that surfaces just‑in‑time context and follow‑ups.
  4. Deeper integrations into Discord / WhatsApp / Telegram and personal calendars.

Long‑term, IRL should feel like an ambient layer: you notice it most when it disappears, not when it’s working.

Who should talk to us

City builders

People already running dinners, clubs, pop‑ups, retreats, or cowork rituals.

Communities with gravity

Discords, Telegram groups, and member‑driven collectives that want more IRL.

Creators & founders

Folks whose work already pulls people into the same city or neighborhood.

Infra partners

Teams building the next generation of social, calendar, or AI infra primitives.

Share how you’re already bringing people together