| https://mlox.org | MLOX Website |
| https://github.com/BusySloths/mlox | Github |
If you're into infra, MLOX is fully open source: backends, plugins, and service definitions are just code. Star the repo, read the internals, and send a PR to add your favorite tool or cloud provider. Let's build a better ML stack together.
Observability by default: using OpenTelemetry in MLOX, you can track traces, logs, and metrics from the start. No more "we'll add monitoring later": the infra is ready when your first model ships.
Want to experiment with LLM serving? With the Docker backend, MLOX can provision LiteLLM + Ollama and wire them into your stack. Quickly test prompt flows, local models, or gateways without learning yet another deployment toolchain.
MLOX lets you deploy models with security and tracking built-in: Log your model to MLflow and then deploy with one command via MLServer. Auth and wiring are handled for you, so you go from “logged run” to “live endpoint” without manual glue scripts.
MLOX has a plugin system designed for extension: add new backends or services by dropping in plugins instead of forking the core. Ideal if you want to integrate your own orchestrator, model store, or custom infra tooling.
Es gibt immer mehr Gründe, nicht *komplett* auf Hyperscaler zu setzen. Hier ist ein neuer: 🚀
In einem aktuellen Konflikt hat der Iran eine Liste mit 29 Tech-Standorten von Amazon, Google, Microsoft & Co. als mögliche Ziele veröffentlicht.
Wenn kritische Infrastruktur zentralisiert ist, wird sie auch geopolitisch relevant.
https://www.golem.de/news/krieg-iran-nennt-29-us-tech-standorte-als-ziele-2603-206419.html
The Docker backend is our most complete environment: spin up Airflow, MLflow, InfluxDB, Postgres, Milvus, LiteLLM, Ollama, and deploy your models via MLServer using the MLOX UI or CLI.
