Gfs

GFS brings Git style version control to databases. It lets a team commit, branch, merge and travel back through database history, so a working copy of realistic data can be spun up for a single task and discarded afterward instead of being shared, reset and queued for by everyone at once. The project comes from Guepard, whose platform creates ephemeral, production shaped databases for AI agent runs, pull requests, continuous integration jobs and experiments. The working pattern is to connect to production in read only mode, take a snapshot, branch it on demand for each agent run or job, then tear the branch down when the work is finished. Supported engines include Postgres, MySQL and MongoDB among several others. The platform can be self hosted inside a customer's own cloud network through Terraform so data never leaves it, with role based access control, PII masking, audit logs that export to a security information and event management system, encryption at rest and in transit, and data residency options in the EU, US and APAC. It is aimed at platform and DevOps teams, release managers and engineering leaders dealing with parallel CI runs, agent workloads and shared staging queues. GFS is released as open source, with version 1.0 published on GitHub. No pricing is published for the hosted platform: there is no plan table and no amounts anywhere on the site, and the calls to action lead to a demo booking or a call with the team. A separate track for startups is advertised without stated terms.