A lot of high-leverage projects (AI or otherwise) involve tying data together from multiple systems of record. This is easy enough when the data is simple and the sources are few, but if you have highly nested data and lots of sources (or you need things like federated pagination and filtering), you have to write a lot of gnarly boilerplate that's brittle and easy to get wrong.
One solution is to import all your data into a central warehouse and just pull it from there. This works, but 1) you need a warehouse, 2) you have an extra copy of the data that can get stale or inconsistent, 3) you need to write and manage pipelines/connectors (or outsource them to a vendor), and 4) you're adding an extra point of failure.
Nerve lets you write GraphQL-style queries that span multiple sources; then it goes out and pulls from whatever source APIs it needs to at query-time - all your source data stays where it is. Nerve has pre-built bindings to external SAAS services, and it's straightforward to hook it into your internal sources as well.
Nerve is made for individual developers or two-pizza teams who:
-Are building agents/internal tools
-Need to deal with messy data strewn across different systems
-Don't have a data team/warehouse at their disposal, (or do, but can't get a slice of their bandwidth)
-Want to get to production as quickly as possible
Everything you see in the demo is shipped and usable, but I'm adding a little polish before I officially launch. In the meantime, if you have a project you'd like to use Nerve on and you want to be a beta user, just drop me a line at mprast@get-nerve.com (it's free! I'll just pop in from time to time to ask you how it's going and what I can improve :) )
If you want to get an email when Nerve is ready from prime-time, you can sign up for the waitlist at get-nerve.com.
Thanks for reading!
(EDIT: Nerve is desktop only! I'll put up a gate on the site saying as much.)