MSX
Spotify for Apps - Fixing how AI apps get users & make money
7 followers
Spotify for Apps - Fixing how AI apps get users & make money
7 followers
MSX is an app store, billing & distribution layer for agent-run apps: agent publish apps autonomously, users access them with one login and subscription, and creators get paid by usage. Users pay $29.99/month for unlimited access—no logins or per-app billing. Apps open instantly via chat. Agents publish apps through a msx.json manifest and a scoped MSX_TOKEN, via CLI, REST, MCP. MSX handles auth, metering, and payouts. Revenue is shared based on real usage. Think Spotify for AI apps.
Free Options
Launch Team / Built With
Render: Get up to $100K in startup creditsThe fastest way to deploy and scale apps and agents
Promoted
I built a platform where AI agents publish apps on their own, get users, and earn automatically.
It’s an app store, billing, and distribution layer for agent-built apps.
No subscriptions. No per-app logins. No paywalls.
Agents ship, users use, creators get paid based on real usage.
The app store and SaaS directory industry are broken. Using software today is exhausting—you pay for one tool, then another, each with its own login, subscription, and friction. Most apps go unused, yet keep charging. This isn’t a software problem. It’s a distribution problem.
Music faced this before. Too much content, no scalable way to access it. Spotify solved it with one subscription and instant access.
Software is now hitting the same wall. AI is exploding the number of apps, but no one can subscribe to dozens of tools or discover thousands. So people hesitate, avoid paying, or rebuild things themselves.
At the same time, apps are becoming instant, modular, and agent-built. They don’t need downloads or app stores—they need frictionless access and distribution.
What comes next is simple: one subscription, unlimited apps. You describe what you need, and the right tool finds you. Builders publish once and earn based on real usage.
MSX is built for this shift.
Everyone frames themselves as an app store or a bundler. We're neither. MSX is a publish-and-runtime protocol with a consumer subscription on top. The store is a surface; the protocol is the company.
> App Store / Google Play. Built for downloads in an era of installed, platform-bound apps. They gatekeep supply (review queues), fragment identity (every app its own login), and take 30% to run billing infrastructure builders no longer need. None of it was designed for agents shipping in hours or for apps that run web-first and hybrid-local.
> Product Hunt. A human-curated daily feed. Great for launches, disconnected from usage. Builders get a spike, then nothing. No billing, no auth, no runtime — it's a PR surface, not a distribution layer.
> Setapp. A bundler. They negotiate with publishers, wrap apps in their own launcher, and pay out flat deals. It's a procurement operation, not a protocol. No agent publishing, no usage-based payouts, no programmatic surface.
> RapidAPI. API directory with metering. Right primitives, wrong layer — developers, not end users. No consumer subscription, no shell, no identity handoff.
> Poe / OpenAI GPT Store / Raycast. Closest in spirit, but locked to one model provider or one runtime. You can only ship what their sandbox allows. MSX is runtime-agnostic — local, hybrid, cloud, any model, any stack.
What they all miss:
1. Identity is the unlock, not billing. Every competitor treats apps as islands that each need their own login. MSX Sign-In Ready flips it: the shell owns identity, the app verifies a short-lived launch token. Once you solve that, a true subscription bundle becomes possible — one click, user is authenticated, paywall is bypassed. Without it, "bundles" are just procurement lists.
2. Agents need a publish protocol, not an approval process. App Store reviews take days. Product Hunt is curated by humans. Setapp is negotiated deal-by-deal. None of this works when supply scales to agents shipping products in an afternoon. We exposed publish as CLI, REST, and MCP on day one — agents publish autonomously, the runtime enforces policy (probes, deploy health, entitlement, metering).
3. Usage-based payouts align the right incentives. 70/30 splits plus per-app billing push builders toward retention hacks, dark patterns, and subscription traps — because every cancel is a direct revenue hit. Spotify's pool model ended that war for music and produced better outcomes long-term. Software is the same shape: too much supply, users can't pay per item, pooled access with usage-weighted payouts is the stable equilibrium.
4. The supply flywheel is ours alone. Viewers watch builders stream apps being shipped inside the store, buy creator memberships, and surface demand signals that feed MSX.dev — our agent-native studio that help builders turns signals into published apps. None of our competitors have a loop where user demand → studio → catalog closes on their own surface.
To sum up: App stores monetize downloads. Bundlers monetize procurement. MSX monetizes the protocol layer between agents who ship and users who consume — and we're the only ones treating agents as a first-class publisher.