Floci is the local cloud for AI coding agents: AWS, Azure, GCP, and OCI emulators that start in milliseconds, so agents build and test against a real-shaped cloud with no credentials and no bill. MIT licensed. No…
49 comments
Many users, (myself included), had need to test cloud stuff locally, but had no tool to do so. Localstack exists, but it started rug pulling the free tier and they didn’t support all the features.
Meanwhile, Floci makes it easy to write your own cloud compatible test suite, then implement the feature to match. I was able to get a good enough coverage on a few features created in a weekend on the $20 tier of Claude to make it useful to me, then others picked up the work to round it out.
It’s really cool to see it grow.
I think we’re in the early days of this, but for extremely verifiable domains like porting from A to B or simulators with an objective oracle, I think it’s just a matter of putting the scaffolding to enable anyone and their Claude to contribute fruitfully.
I’m hoping we can do a similar thing with shared search for vulns, though it’s a bit harder. But if eg repos give an AGENTS.md with instructions on the bar for a good reproducer, then you could start to see more helpful federation patterns (instead of drive-by CVEs which are a net drain on contributors).
Different OSS project management skills required, but I am optimistic that we can do better at onboarding non-experts to federate work via their Claude subs. Almost like Folding at Home.
1. https://github.com/vercel-labs/emulate
for modelling and testing a CI software i was working on, i had to extend their github emulator in several places for example. but it was great to already have a working base.
I’ve definitely reached for these local cloud tools from time to time, but in a well-architected piece of software, you aren’t getting that much value from them. Either you don’t have a hard cloud dependency, or you want a high fidelity test of a specific cloud dependency.
Reached the same conclusion. Plus it's a fair amount of effort to set them up and maintain.
I think if I was doing evals on the way an agent interacts with a service (e.g. Slack) the digital twin could make sense, but for rapid iteration or testing, it might not be what you want.
Read the full thread on Hacker News →
Related stories
- Hacker News · 1 points · 5 days ago
- Hacker News · 1 points · 3 days ago
- Hacker News · 1 points · 8 days ago
- Hacker News · 1 points · 6 days ago
- Hacker News · 1 points · 10 days ago
- Hacker News · 1 points · 10 days ago