Muse AI Connectors: The Next App Store Moment?

Meta’s decision to let developers submit Muse connectors could become an “app store moment,” but the analogy is a possibility, not evidence that a large market already exists. At the time of recording, Muse was the App Store’s number one app. Apple had paid developers $1 billion from apps and in-app purchases by June 2010, less than two years after its App Store opened. Whether Muse connectors reach comparable scale depends on how many people use Muse for the specific tasks connectors can handle.

Meta offers a submission process and a directory of approved integrations, which could expose small and medium-sized businesses to customers and help early entrants learn what users need. But the size of Muse’s user base, the share of connectors Meta will approve, and whether people will find and repeatedly use directory listings remain unknown. A listing may help with exposure, but the discussion cautions against treating directory placement or a Meta feature as a dependable customer-acquisition plan.

Connectors can retrieve information from business systems, present an offer, and, with user approval, complete a booking and return confirmation; they do not take over the business’s operational responsibilities. A venue still has to maintain accurate availability and equipment and help customers access it. Companies with existing APIs may be able to build on capabilities they already provide. Duffel says its Muse integration supports flight searches and booking management, and reports more than $1 billion in annual transaction value—booking value across the company, not Duffel revenue.

Two proposed connectors illustrate the need to validate both customer demand and a paying buyer. A restaurant-opening service could identify nearby restaurants expected to open and sell suppliers verified, timely leads with public business contacts, but plans can go stale and an opening does not prove a restaurant needs a particular supplier. The suggested test is one city and one supplier, asking whether the supplier will pay regularly; 100 customers at $99 per month would generate $9,900 in monthly recurring revenue before costs. A home-repair connector could match an appliance model with providers able to do the job the next day, then pass approved details or book where supported. Providers might pay $100 for a qualified introduction or confirmed booking, but the service must deliver leads they want. Testing one repair type in one location can reveal provider capacity and local demand; homeowners’ occasional needs make steady provider work part of the challenge.

A paddle connector could show nearby games at a player’s level, prices, and bookable courts or matches, but it depends on clubs sharing availability and accepting reservations. Playtomic already offers court and match booking, so a possible gap is availability at local clubs that players currently check separately. The discussion expects Muse may become more social over the next six to 12 months, given Meta’s social expertise, but shared bookings are only a potential growth loop: invitations must demonstrably bring in customers. A family-dinner connector could plan meals around available time and ingredients, create a shopping list, and make it shoppable through Instacart’s developer tools. A subscription is worth testing with families who use it weekly, then improving plans from their meals and feedback. The suggestion that companies such as Instacart might acquire connectors in the next couple of years is speculation conditional on market growth.

For an initial build, the recommendation is to define one customer task, interview reachable customers about a recent instance, and automate only a small part of the job. Availability checking and a quote that states the full price and how long it remains valid can precede booking. Coding agents such as Claude Code or Codex may turn API documentation and a plain-language request into a small service, but builders still need to inspect and understand the result, host it where Muse can reach it, and restrict users to permitted information and actions. Meta documents a custom-connector option to try a supported integration using API information before applying for a public listing. A submission account described by StackTrees’ founder requested product information, use examples, documentation, and support details, and offered an API or existing MCP server route; approval is required, and applicants should check the current form because the program is new. Before submission, test already-booked times, expired quotes, duplicate requests, and booking changes. For customer acquisition, the discussion favors relevant creators, marketplaces, and partners over reliance on directory discovery; Ticketmaster says eligible events can appear through its connector without each organizer doing additional integration work.