Self-hosted AI infrastructure, unpacked.
Run open models on your own GPUs with the developer experience of a managed API — gateway, keys, observability, and a playground included.
Inference
OpenAI-compatible inference behind one endpoint. Issue a key, point your SDK at it, done.
api.inference.unpack.sh/v1
Blog
Mental models and field notes on engineering and AI infrastructure.
Lab
Experiments in fluent engineering — prototypes before they become products.