mvp · early pilot · voice-native

Speak it. It's booked.

CatchUp is a real-time voice agent for local services. One natural sentence becomes a confirmed booking with a real merchant — no app tour, no forms, no callbacks.

Built on a single real-time voice model — not a transcribe → think → speak pipeline. Starting with beauty and wellness.

sub-second latency live merchant calendars no app to install
live · agent session
00:03.9
user · livecontinuous stream

"haircut near my office, tomorrow after six…"

agent reasoning in-flight · tools firing in parallel

+120ms
intent.detected
haircut · ≥18:00
+260ms
merchants.shortlist
12 salons · 1.5km
+410ms
calendar.window
Wed · 17:30–20:30
+540ms
slot.lock
North & Co. · 18:15
user interrupts: "actually, with Marco"
agent adapts in-flight · slot re-locks · no restart, no turn
Booked · North & Co. · Marco · 18:15
user heard “you're booked” in the same call · no handoff, no callback
Real-time by design

Real-time, end to end.

Most voice assistants are three slow systems stitched together: transcribe what you said, ask a model, then read the answer aloud. CatchUp is built on a single real-time voice model — it hears, thinks, and speaks as one continuous flow.

One continuous conversation

Speech in, speech out, in the same stream. No turn-by-turn ping-pong between separate transcribe, think, and speak services.

Sub-second responsiveness

CatchUp listens while you speak, reasons in flight, and replies the moment you stop, the way a good receptionist would.

Interruptible and adaptive

Change your mind mid-sentence, ask a follow-up, switch language. The conversation adapts in real time, not after a delay.

The old way
speech → [ STT ] → text → [ LLM ] → text → [ TTS ] → speech

Three handoffs. Latency stacks. Tone, pauses, and intent get lost between boxes.

CatchUp
speech ⇄ [ real-time voice agent ] ⇄ speech

One model, one stream. It hears your tone, holds context, talks to merchant systems live, and books while you're still talking.

Merchant side, in real time

The booking workflow runs while the user is still talking.

As the conversation unfolds, CatchUp is already querying merchant calendars, holding slots, and pinging the front desk. By the time the user stops speaking, the booking is ready to confirm — not queued for a callback.

user · live audio stream
streaming
  1. t = 0.4s
    “I need a haircut…”
  2. t = 1.6s
    “…near the office, tomorrow…”
  3. t = 3.1s
    “…after six, with Marco if he's free.”

no “processing…” pause · no callback · no form

merchant ops · north & co.
live workflow
triggered at t = 0.4sreal-time
+120msintent.detectedservice_class: hair
+260msmerchants.shortlist12 salons in 2km radius
triggered at t = 1.6sreal-time
+90msgeo.resolvedanchor: work · radius 1.5km
+180mscalendar.windowWed · 17:30–20:30
triggered at t = 3.1sreal-time
+70msstylist.matchMarco @ North & Co.
+140msslot.locked18:15 · held 90s
+210msmerchant.notifiedpush → front desk tablet
Front desk sees the booking before the user finishes the sentence.
one-tap confirm → user hears “you’re booked” in the same call
The problem

Local services have a coordination problem.

A user may need a haircut near work tomorrow after 6. A salon nearby may have an empty stylist slot at exactly that time. But today, that match is hidden across Google, marketplaces, Instagram, WhatsApp, phone calls, and disconnected booking tools.

CatchUp connects both sides: users express intent in natural language, merchants expose real availability and constraints, and the system turns the match into a confirmed booking.

Users bring demand

Natural-language requests like “book a haircut near work tomorrow after 6.”

Businesses bring capacity

Available staff time, services, prices, constraints, and open slots.

CatchUp creates the transaction

Intent is matched with availability, confirmed, reminded, and followed up.

MVP booking flow

From spoken request to confirmed booking.

Capture, Resolve, Match, Commit. Four stages that turn a natural-language request into a real transaction with a real merchant.

01

Capture

The user asks in natural language by voice or text.

02

Resolve

CatchUp understands service type, location, timing, constraints, and preferences.

03

Match

The system compares the request with merchant availability, services, and rules.

04

Commit

CatchUp moves from suggestion to booking confirmation and follow-up.

For merchants

Turn unused time into booked revenue.

Beauty and wellness businesses sell perishable time. An empty slot today cannot be sold tomorrow. CatchUp helps merchants expose availability, automate booking workflows, reduce no-shows, and bring customers back.

Capture demand

Voice and text requests routed straight to your availability.

Fill open slots

Match unused time with relevant nearby demand.

Automate reminders

Confirmations and reminders that reduce no-shows.

Cut manual follow-up

Fewer missed calls, fewer back-and-forth messages.

Rebook & retain

Targeted rebookings and campaigns to your customer base.

Join the merchant pilot

We're onboarding a small group of beauty and wellness businesses for the early pilot. Light setup, real bookings.

Apply to the pilot
For users

Ask once, get the booking done.

Users do not want another app to browse. They want the job done. CatchUp lets them ask naturally and get real options, real availability, and a confirmed booking.

  • Search by intent, not filters
    Speak or type the request the way you'd say it out loud.
  • Real options, real availability
    Matched on time, location, and your preferences, not a stale listing.
  • Book without app-hopping
    No more bouncing between Google, Instagram, and a booking tool.
  • One thread for the whole booking
    Confirmations, reminders, and follow-ups stay in one flow.
user request
Book a haircut near my office tomorrow after 6.
user request
Find a massage nearby this evening.
user request
Rebook my usual cut, next Thursday.
confirmed · Booked at North & Co., 18:15. Reminder set.
Why CatchUp

Not a chatbot. Not a directory.

General AI can suggest places. Directories can list businesses. Booking platforms can show some availability. CatchUp is different because it connects user intent with merchant-side workflows and turns that match into a real transaction.

Chatbot

Suggests options.

Directory

Lists businesses.

Booking app

Shows inventory.

CatchUp

Matches intent with capacity and completes the booking.

Wedge

Beauty and wellness first. Local services next.

We start with beauty and wellness because appointments are frequent, capacity is perishable, and many businesses still rely on fragmented tools. The broader opportunity is to become the transaction layer for bookable local services.

  • Frequent, repeat appointments
  • Perishable, time-sensitive capacity
  • Fragmented merchant tooling today
  • Voice-friendly consumer behavior
Thesis

The transaction layer for bookable local services. Built one category at a time, starting now.

mvp pilot · early access

Help us build the transaction layer for local services.

We are onboarding early beauty and wellness merchants and early users for the MVP pilot. Tell us who you are — we'll route you accordingly.

Join the merchant pilotRequest user access
I am a

No spam. We'll only email about the CatchUp pilot.

FAQ

Questions, answered.