Stripe Projects: Simplify App Development with Instant Service Provisioning

This title was summarized by AI from the post below.
View profile for Patrick Collison
Patrick Collison Patrick Collison is an Influencer

When Andrej Karpathy built MenuGen (https://lnkd.in/gQzHkipf), he said: "Vibe coding menugen was exhilarating and fun escapade as a local demo, but a bit of a painful slog as a deployed, real app. Building a modern app is a bit like assembling IKEA future. There are all these services, docs, API keys, configurations, dev/prod deployments, team and security features, rate limits, pricing tiers." We've all run into this issue when building with agents: you have to scurry off to establish accounts, clicking things in the browser as though it's the antediluvian days of 2023, in order to unblock its superintelligent progress. So we decided to build Stripe Projects to help agents instantly provision services from the CLI. For example, simply run: $ stripe projects add posthog/analytics And it'll create a PostHog account, get an API key, and (as needed) set up billing. Projects just launched as a developer preview and we've been inundated by the response. You can register for access (we'll make it available to everyone soon) at https://projects.dev. We're also rolling out support for many new providers over the coming weeks. (Get in touch if you'd like to make your service available.)

Interesting direction. As agents begin provisioning execution environments directly, infrastructure increasingly becomes machine-addressable rather than human-configured. One implication is that decision layers may scale faster than mechanisms for verifying the integrity of the financial objects those agents interact with. When financial workflows become composable at machine speed, confirming that underlying obligations remain coherent and unique across systems becomes as important as enabling agents to provision the rails themselves. Execution abstraction is advancing rapidly — trust primitives may need to evolve in parallel.

This resonates a lot. We’re seeing a similar problem one layer up, not provisioning services, but determining what should be allowed to transact through them. Building OATH to resolve product eligibility + compliance in real time (HSA/FSA, medical, regulated commerce) — basically a decision layer sitting alongside the infra you’re describing. Curious if you see this becoming part of the developer control plane as well.

The protocol you are co designing here is the real story. We are moving away from manual service integration toward a world where the infrastructure itself is deterministic. For those of us building in regulated markets where audit trails and credential security are nonnegotiable this level of automation is a massive relief. It changes the conversation from how we connect to what we are actually building.

Very real problem being addressed here. As systems become more composable, the bottleneck shifts from capability to orchestration and governance of dependencies. Abstractions like this can significantly improve developer velocity, but over time, visibility into costs, controls, and third-party risk will become equally critical. Interesting to see how platforms like Stripe balance speed with governance as this evolves.

Karpathy's IKEA furniture analogy captures the problem but frames it too charitably. IKEA furniture at least has one set of instructions and one assembly sequence. Modern app deployment is more like assembling furniture where the instructions reference three other furniture brands, some of which updated their parts since the instructions were written, and the load-bearing pieces are managed by services with their own outage windows. The deeper issue his complaint reveals: the industry hasn't built good abstraction layers between 'this idea works locally' and 'this idea survives contact with production.' The complexity isn't accidental. It's the accumulated technical debt of a generation of tools built for infrastructure engineers, now used by people who just want to ship a product.

Like
Reply

This feels like a natural step in the evolution of agent workflows. The real bottleneck has never been intelligence - it’s orchestration. Anything that removes friction between idea and deployed infrastructure will quietly redefine how fast teams can actually build.

This solves one of the most painful parts of building real apps. I recently vibe coded something and payments were so painful, and I kept wondering about security. But Andrej's full list goes beyond what Stripe owns. The real unlock is integrating this with agents like Codex or Claude that can orchestrate provisioning, configuration, and deployment across providers in one flow. That's when vibe coding can go from demo to production!

I like the direction here. The bottleneck isn’t intelligence anymore, it’s setup… if agents can provision their own stack, that’s a pretty meaningful unlock.

Most people will read this as workflow convenience, but the expensive consequence is that service choice and system shape start getting decided wherever the agent can act fastest.

The real bottleneck has always been the surrounding setup rather than the build itself. Eliminating that friction accelerates the transition from ideas to products, and it will be interesting to see the expansion of provider integrations.

See more comments

To view or add a comment, sign in

Explore content categories