Show HN: Nerve – a bottom-up, network-based approach to universal integration
get-nerve.com
Nerve was made to help software developers build and maintain integrations. This problem is fairly infamous by now so I won't belabor the point; suffice it to say that writing integrations sucks, it's only getting worse, and no one so far has improved the experience enough to dominate the market. Nerve takes a novel approach to software integration, but it builds on the concept of a Unified API. In a nutshell, a Unified API is a 1:n API which presents 1 endpoint as a facade for n different sources. It normalizes everything into a single data model and has a single way of doing authentication, rate limiting, etc. If (for example) you were building a SaaS app and wanted to hook into your customers' CRM, you could develop against one Unified API instead of building integrations for each of the 30 CRMs your customer could be using.
Unified APIs are starting to get some broader recognition, and accordingly the space has a lot of recent entrants (Merge, Nango, and Apideck come to mind, but there are a ton of others). These are all great companies (and well-loved), but I consider them more tech-enabled service businesses than pure software plays, since the way they make integrations easier is not by making the act of integrating more efficient, but by writing and maintaining integrations on the customer's behalf. They're essentially grinding the problem down by throwing bodies at it (in Nango's case these are open-source bodies, but bodies nonetheless).
As software continues to proliferate and end users continue to expect more compatibility and ease-of-use, I believe it's going to be tough for these Unified API companies to satisfy developers' escalating demands for broader and deeper API coverage while both maintaining the level of customer service they're currently known for and ensuring their margins don't collapse. I think that, unfortunately, most Unified APIs are going to end up leaving us the way we started - with a patchy and frustrating developer experience.
Nerve is my implementation of what I believe to be a more scalable and tech-first approach to large-scale software integration. Instead of being a Unified API per se, Nerve is what I call a Schema Network - like the name implies, the strategy is to take the point-to-point integrations that people are already writing and tie them together into a globally connected network (think of apps as nodes, and integrations as edges. The basic idea is that you can auto-derive integrations between any two apps that have a path between them on the network - of course, making this idea work smoothly in production involves a lot of subtlety!). Nerve uses its Schema Network to present developers with a unified interface to many apps at once. In fact, a Schema Network is not that different from a Unified API from the perspective of the user - what makes Nerve unique is that its unified schema evolves from the bottom up out of local integrations made by local actors, instead of being managed from the top down by a central team. Removing the bottleneck of centralized control allows Schema Networks to tie more services together than is possible with Unified APIs, and at a much higher level of detail.
The Internet let all of the world's programs talk to one another. I built Nerve because I strongly believe that a Schema Network, or something like it, is the only feasible way for all of the world's programs to understand each other.
If you would like to join the waitlist for the private beta, there's a short form you can fill out [1] (and take a look at [2] if you'd like more details about the beta). And of course, all questions, thoughts, and feedback are always very much appreciated! Thanks so much for taking a look!
[1] https://2vkwx88tksj.typeform.com/to/emg5xLFC [2] https://gist.github.com/mprast/62a5afd5db5075584756e123135b9...