Floci is a free, open-source multi-cloud emulator suite built on Quarkus native image, created by Hector Ventura. Unlike mocking or simulation tools, Floci provides true emulation — storing and returning real data — for AWS, Azure, GCP, and Oracle OCI, each running in its own container. Compiled to a GraalVM native image via Quarkus, it achieves 24 ms startup time and 13 MB idle memory. It integrates with Quarkus Dev Services as a drop-in replacement for LocalStack (which archived its community edition shortly after Floci launched). Floci offers four storage modes (memory, persisted, hybrid, write-ahead log) and is increasingly used for AI agent sandboxing, where isolated, disposable cloud environments per agent eliminate risks of LLM operations touching real cloud accounts. Upcoming features include CloudFormation expansion and 'Floci Pods' — portable snapshots of complete environment state for reproducible bug reports.
Questions this post answers
What are the startup time and memory usage of Floci?
Floci starts in 24 ms and uses 13 MB of idle memory. These numbers come from compiling the entire emulator to a GraalVM native image via Quarkus, the same native compilation pipeline that powers Quarkus's benchmark performance. The lean footprint makes it practical for CI/CD pipelines and AI agent sandboxing where many isolated instances may run in parallel. Teams running Floci in CI or agent workflows track performance regressions and new releases on daily.dev.
How do I redirect the AWS CLI to a local Floci instance?
Set the AWS_ENDPOINT_URL environment variable to http://localhost:4566, use any placeholder credentials (Floci does not validate them), and rely on Floci's built-in DNS alias — *.floci.io resolves to 127.0.0.1 — so both path-style and virtual-hosted-style S3 URLs resolve locally without editing /etc/hosts. After those three steps, standard commands like aws s3 ls and aws s3 mb work as they would against a real account. Developers switching local AWS tooling from LocalStack to Floci find the migration details on daily.dev.
What storage modes does Floci support for local cloud emulation?
Floci offers four storage modes: Memory (fastest, data lost on container stop, ideal for AI agent sandboxes), Persisted (every write goes to disk immediately, good for deterministic test fixtures), Hybrid (writes go to memory and flush to disk asynchronously, inspired by PostgreSQL's write-ahead log), and Write-ahead-logging (tailored for write-heavy workloads). Persistence was a first-class design goal from the start. Engineers building reproducible test environments with local cloud emulators follow Floci's roadmap on daily.dev.