Roadmap

Trip Creation & Invites shipped

The core loop starts here: one person creates a trip and shares a link. That's it. No app to download, no account required for the people joining. The organizer picks some potential destinations and a rough date window, and gets a unique invite link like crab.travel/to/5RSfhxSu that they text to the group.

When someone opens the link, they see the trip details and fill out their preferences — budget range, interests (craft beer, hiking, live music, etc.), travel style (adventure, relaxation, foodie), accommodation preference (hotel vs. Airbnb vs. flexible), dietary needs, and any mobility considerations. They can do this without creating an account, just enter their name and email. If they want to sign in with Google for a richer experience (calendar sync, profile that carries across trips), they can.

The organizer sees a dashboard of everyone who's joined, their preferences, and who's still outstanding. The goal is to get from "hey let's go somewhere" to "everyone's weighed in" with a single link and zero friction.

What's built:

Voting & Availability shipped

The hardest part of group travel isn't finding a good hotel. It's getting 12 people to agree on where and when. That's a coordination problem, not a search problem. So we built voting and availability as first-class features, not afterthoughts.

Destination voting uses rank ordering: if there are 4 destination options, each member ranks them #1 through #4. This gives the organizer way more signal than a simple thumbs-up/thumbs-down. You can see that Nashville is everyone's #1 or #2, while Vegas is either #1 or #4 (polarizing). The vote counts show in real time on the destination tabs so people can see consensus forming.

Date availability uses a visual calendar where each member marks their dates as "Ideal" (bright green), "If needed" (amber), or "Can't go" (red). Everyone's availability is overlaid on a shared group calendar with names showing on each date. The organizer can immediately see which week has the most green and the least red. No back-and-forth texting required.

The date picker is inline. You click a start date and end date on the same calendar view, not separate popups. The trip progresses through stages (Voting, Planning, Booked) that the coordinator controls from their dashboard, so everyone knows where things stand.

What's built:

CrabAI Destination Research shipped

Once destinations are on the table, CrabAI goes to work. For each destination, it generates what we call a "destination card", think of it like a Pinterest board specifically curated for your group. It's not generic travel advice; it's filtered through your group's actual preferences, budget, and interests.

Each card includes: 3-4 stays (hotels, Airbnbs, or resorts with price ranges and neighborhood context), 6-8 activities (real places with real names, not "visit a museum" but "National Museum of African American Music on 5th Ave"), 4-5 restaurants and bars (matched to the group's foodie vs. casual vibe), and 3-5 upcoming events (concerts, festivals, sports games happening during your travel window). Every card also shows an estimated total cost per person for a 5-day trip and a compatibility score from 0-100.

The "group vibes" feature is particularly powerful: if the coordinator tags the trip as "tailgating" or "bachelor party" or "wine country," CrabAI filters ALL recommendations to match that vibe. A tailgating trip to Nashville won't suggest art galleries. It'll find sports bars near the stadium, BBQ joints, and Titans/Predators game schedules.

There's also a "Suggest for us" button that lets CrabAI propose 5 completely new destinations based on the group's collective profile, useful when nobody has a strong opinion and you want to be surprised.

What's built:

Travel Search & Deals shipped

CrabAI recommendations are great for discovering what to do somewhere, but when it comes to flights and hotels, people want real prices they can actually book. So behind the destination cards, we run actual travel searches across multiple providers simultaneously.

The search engine fans out to four adapters in parallel: Duffel (flexible flight bookings), Travelpayouts (global flight deals with affiliate revenue), LiteAPI (multi-source hotel and activity aggregation), and Viator (tours and experiences. The world's largest activities marketplace). Results stream in via Server-Sent Events, so prices start appearing within seconds of triggering a search rather than making you wait for all providers to respond.

Results are deduplicated across providers (same flight from two sources → show the cheapest), stored in the database with timestamps, and linked to the trip. Each result includes a deep link to the booking partner, so when someone's ready to buy, they click through directly to Expedia, Booking.com, or Viator to complete the purchase. We earn affiliate commission on those bookings. That's the revenue model for this layer.

There's also a global deals dashboard that runs daily via cron job, checking for the cheapest flights from 14 major airport hubs. This feeds a "hot deals" view that surfaces opportunities even before a trip is fully planned. Sometimes the best trip is the one where the flight is $89.

What's built:

Interactive Demo Mode shipped

You shouldn't have to sign up to see what a trip looks like. The demo system lets anyone experience the full platform as "Judy Tunaboat". A fictional member who's part of every trip. Click "See a demo trip" on the homepage and you're instantly inside a fully booked Scottsdale trip with 12 members, real flight prices, hotel bookings, and CrabAI-generated group chat.

A stage switcher lets you see what Voting, Planning, and Booked stages look like: three different trips at three different phases of the planning process. Navigate away from the demo to any regular page and your real account restores automatically. No login required, no friction.

What's built:

Chat & Notifications shipped

Every group trip lives and dies by communication. The problem with using a group text is that it mixes trip logistics ("has anyone booked the Airbnb?") with regular life ("lol check out this meme"). And once the thread gets long enough, people stop reading it entirely. Important decisions get buried.

crab.travel has a dedicated trip chat that lives on the plan page. It's collapsible and sits at the top of the page so it's always accessible but not in the way. Messages support threaded replies so side conversations don't derail the main thread. If you're signed in, your name shows automatically; anonymous members use their display name.

For notifications, each user chooses their preferred channel (email, SMS, or both) and frequency (real-time, daily digest, weekly digest, or off), separately for chat messages and trip updates (votes, new members, stage changes). The idea is that the organizer who wants to know everything immediately can get real-time pings, while the casual member who just wants a weekly summary can get that instead.

Email notifications are live now. SMS is built and ready to go. We're waiting on Twilio's A2P campaign approval (resubmitted March 20, 2026) before we can send texts. Once that clears, chat messages and trip updates will push to phone via text for anyone who opts in.

What's built:

Price Drop Alerts shipped

We already track every price we see. The price_history table logs every flight and hotel price with timestamps, so we know exactly how prices move over time for any route or property. The infrastructure is there. We just need to connect the dots and tell people when something gets cheaper.

Here's how it would work: once a trip locks in a destination and date range, we run the search adapters on a recurring schedule (probably daily or every few hours) instead of just on-demand. Each new result gets compared against the last recorded price for that route or hotel. If a flight drops by more than 10% or falls below a threshold the user set, we fire off an email or SMS: "SEA → BNA dropped from $247 to $189 on June 12. Book now?" with a direct link to the booking page.

The stretch goal is a "best time to book" indicator based on price trend analysis: are prices trending down (wait) or starting to climb (book now)? Airlines and OTAs have this data but they use it to maximize their revenue. We'd use it to maximize the traveler's savings.

Why this matters: Most people book flights based on vibes ("this seems like a good price") because they don't have time to check every day. Price monitoring is the kind of thing that's trivial for software to do and painful for humans to do. It's a perfect automation target and a genuine "wow, this saved me money" moment that builds trust in the platform.

Planned:

Per-Person Cost Breakdown shipped

This is the spreadsheet killer. Right now, the person organizing a group trip inevitably ends up maintaining a Google Sheet with columns for each person, rows for each expense, and formulas that break every time someone changes something. It's the single most tedious part of group travel and it persists long after the trip is over ("hey, you still owe me for the Airbnb deposit").

The basic version is simple: each trip gets a cost section where the organizer (or any member) can add line items: "Airbnb deposit: $2,400 paid by Adam" or "Kayak rental: $180, split 6 ways." Each item gets tagged as shared (split evenly), per-person (individual), or custom-split. The platform computes a running total per person and shows a clean dashboard: who's paid what, who owes what, and the net balance between any two people.

We're not building a payment processor. That's Venmo and Zelle's job. What we'd do is generate deep links: when you see "You owe Adam $247," there's a button that opens Venmo pre-filled with the right amount and a note. Settlement in two taps.

Why this matters: Money is the #1 source of friction in group travel. Not because people are cheap, but because nobody has visibility. When everyone can see the same number in the same place, arguments disappear. This feature alone could be the reason someone chooses crab.travel over a group text.

Planned:

Day-by-Day Itinerary in progress

Right now, the destination card shows a flat board of pins (stays, activities, restaurants, events) all in one view. That's great for the discovery and voting phase ("ooh, that BBQ place looks amazing"). But once the trip is locked in, people want a schedule: what are we doing Thursday? What about Friday morning?

The itinerary builder would take the pins from the destination card and let the organizer (or anyone) drag them into day-by-day time slots. Thursday morning: arrive, check into hotel. Thursday afternoon: kayaking. Thursday evening: dinner at Martin's BBQ, then Broadway honky-tonks. Friday morning: Parthenon replica, then brunch. The whole group sees the same timeline and can suggest swaps or additions.

CrabAI here would suggest optimal ordering based on geography (don't put a morning activity 40 minutes from the afternoon one), operating hours (museum closes at 5, don't schedule it for 6pm), and group preferences (morning people vs. night owls based on what they said in their profiles). It wouldn't force anything, just surface suggestions like "swap these two so you're not driving across town twice."

Why this matters: The itinerary is what makes a trip feel real. It's the difference between "we're going to Nashville" and "we're going to Nashville and here's exactly what we're doing." It also reduces the on-the-ground decision fatigue that kills the vibe on day 2 of every group trip ("I don't know, what do YOU want to do?").

Planned:

Room Assignments & Flight Tracking next

This one comes directly from watching how real group organizers work. When you're coordinating 15+ people, you end up maintaining a document that tracks: who's rooming with who (especially for Airbnbs with multiple bedrooms), what everyone's flight details are (so you know who's arriving when and whether you need to coordinate airport pickups), and a master summary you can send to the group.

Room assignments: The organizer sees a list of all booked accommodations (say, a 5-bedroom Airbnb and a 2-room hotel block) and can drag members into rooms. Visual layout: Room 1 = Adam + Pete, Room 2 = Tim + Gene, etc. Members see their assignment on the trip page. If someone wants to swap, they request it and the organizer approves.

Flight tracking: Each member enters their flight info, either a confirmation code (we'd pull the details from the airline) or just airline + flight number + date. The trip page shows a "Who's arriving when" timeline: three people land at 2pm, two more at 5pm, one at 9pm. The organizer can see at a glance whether they need one airport run or three.

The PDF export is the cherry on top: a clean, printable summary of the entire trip: who's in which room, everyone's flight details, the itinerary, cost breakdown, and key contact info. It's literally the PDF that organizers are already making by hand, but auto-generated from the trip data. Print it, email it, or save it to your phone.

Planned:

SMS Voting & Inbound Commands next

Here's the reality of group travel: in a group of 15, maybe 5 will actually open the link and use the site. The other 10 will glance at it on their phone, think "I'll do this later," and never come back. They're not being rude. They're just busy and the activation energy of opening a browser, scrolling through options, and clicking buttons is too high.

SMS voting meets those people where they are. The organizer triggers a vote, and each member gets a text: "Time to vote! Where should we go? Reply 1 for Nashville, 2 for Phoenix, 3 for Austin." The member replies "1" and they're done. Their vote is recorded on the site, the tally updates in real time, and they never had to open a browser.

Beyond voting, inbound SMS could handle other lightweight actions: "Reply YES to confirm you're attending" or "Reply with your flight number" or even forwarding chat messages as texts with the ability to reply back into the chat. The webhook infrastructure for inbound SMS already exists. We just need to build the command parser on top of it.

Why this matters: Participation rate is everything for group coordination. A vote with 5 of 15 responses isn't useful. A vote with 13 of 15 is. SMS dramatically lowers the bar for participation, especially for the less tech-savvy members of any group (parents, older friends, the buddy who still uses a flip phone).

Planned:

Email Digest for Lurkers next

Most trip members fall into one of three categories: the organizer (checks the site constantly), the engaged few (check it when they get a notification), and the lurkers (never check it unless something forces them to). The lurkers aren't disinterested — they just need a different cadence. Hitting them with real-time notifications creates alert fatigue. But a weekly email that says "here's what happened this week with the Nashville trip" keeps them in the loop without overwhelming them.

The digest would be a clean, well-designed email summarizing trip activity since the last one: new votes and current standings, price changes ("flights to Nashville dropped $40 this week"), new members who joined, chat highlights, and any decisions the organizer made. Each section would have a one-click action: "Vote now," "View prices," "See full chat." The goal is that even if a lurker never visits the site, they still feel informed and can participate in 30 seconds from their inbox.

Smart timing matters here: don't send a digest if nothing happened. Nobody wants an email that says "no updates this week." The system should only fire when there's actual news worth sharing. A new vote, a price change, a stage transition, or a message that got 3+ replies.

Planned:

Shareable Trip Cards future

Right now, when someone shares a crab.travel invite link in iMessage or a Slack channel, it shows up as a plain URL. Every invite link is a potential viral moment. Someone sees a cool-looking trip card in a group chat and thinks "I want to plan something like that too."

A shareable trip card would generate a dynamic OG image (Open Graph. The preview that shows when you paste a link) with the destination photo, trip name, date range, member count, and maybe a fun stat like "Nashville is winning with 8 votes." It would look polished in iMessage, WhatsApp, Slack, Twitter, Instagram stories — anywhere links get shared.

The technical approach: generate the image server-side on first request (using something like Puppeteer or a headless Chrome endpoint), cache it, and serve it via the OG meta tags. The image regenerates when trip details change materially (new destination locked, stage change). This is a growth multiplier disguised as a feature — every shared link becomes a mini-ad for the platform.

Planned:

Group Booking & Negotiated Rates future

When 15 people need hotel rooms in the same city on the same dates, that's leverage. Hotels offer group rates starting at 10 rooms, sometimes as low as 5. The discount can be 15-30% off rack rate, plus perks like a free meeting room, complimentary breakfast, or late checkout. But most people don't know this exists, and even if they do, they don't want to call the hotel's group sales department and negotiate.

crab.travel would handle this automatically: once a trip is locked and we know the headcount, destination, and dates, we could query hotel group booking APIs or even reach out to properties on behalf of the group. The organizer sees: "Marriott Downtown Nashville: $159/night (group rate, normally $219). Minimum 10 rooms. 12 of your 15 members need rooms. Book as a block?"

Same concept applies to activities (group pricing on tours, rentals, experiences) and potentially flights (group booking with airlines for 10+ passengers, though this is harder). The revenue model is clean: we take a commission on group bookings, and the travelers still pay less than they would individually. Everyone wins.

Planned:

Retiree Life Concierge future

This is the second major use case and it shares 80% of the infrastructure with group trips. The insight: retirees have disposable income, time, and a deep desire to stay connected to family and active in life, but often lack the structure, tech savvy, or activation energy to make it happen. Nobody is serving this demographic beyond financial planning. We'd cover everything else.

The onboarding would be modeled after a dating profile — familiar, non-intimidating, even fun. Tell us about you (lifestyle pace, personality, home base). What lights you up? (tap interest tiles — golf, cooking, history, music, gardening, travel, theatre, pickleball). Your family (add grandkids by name/age/city, invite adult children to connect calendars via text link). Travel style (budget, distance comfort, hotel vs. rental, frequency, physical considerations).

CrabAI would synthesize everything into a monthly "Life Plan" — events near them that match their interests, trip ideas matched to their profile, upcoming family moments flagged from connected calendars (grandkid's soccer game, daughter's birthday). Delivered by a human advisor in a warm monthly touchpoint — CrabAI does the work, and a real person does the relationship.

Distribution channel: Financial advisors white-label it as a premium client benefit. Advisors already serve retirees and are always looking for ways to add value beyond portfolio management. They'd offer "life planning" as a premium add-on, we'd get distribution without direct customer acquisition cost. AARP, AAA, and senior living communities are also natural partners.

Planned:

Local Event Marketplace future

This is a two-sided marketplace that serves both the traveler and the local business. Right now, when CrabAI recommends "check out the Ryman Auditorium" or "there's a food festival downtown," it's pulling from general knowledge. But what if local businesses could submit their own events and reach a hyper-targeted audience of groups who are already planning to visit their city?

The submission side is deliberately low-friction: paste a Facebook event URL, forward an email, upload a PDF of a flyer, or fill out a simple form. CrabAI normalizes everything into structured data — auto-tagging by category (music, food, sports, outdoor, cultural), physical demand level (easy walk vs. strenuous hike), price range, duration, indoor/outdoor, family-friendly, and accessibility. A local BBQ restaurant running a weekend cookoff just pastes their Facebook link and it's in the system, tagged and ready to match with incoming groups.

Revenue model: Basic submission is free. Paid featured placement gets your event promoted to relevant groups visiting your city. A "verified partner" badge (monthly fee) signals trust and gets priority in CrabAI recommendations. The audience is incredibly valuable. These aren't random tourists browsing TripAdvisor, they're groups of 10-20 people who have already committed to visiting and are actively deciding what to do.

Planned: