Modelplane helps teams run their own AI inference fleets
Modelplane is an open source for running GPU clusters as one fleet. It can combine resources across s, neos, and s.
It works above serving tools that teams may already use, including vLLM, SGLang, and TensorRT-LLM. Its job is the fleet layer across more than one cluster, not just the serving layer inside a single cluster.
It handles model placement, request routing, autoscaling, and weight caching. The project is aimed at teams that plan to run themselves instead of relying only on APIs.
Key points
- Modelplane turns GPU clusters across different environments into one fleet.
- It supports cloud, neocloud, and on-prem resources.
- It builds on existing serving tools such as vLLM, SGLang, and TensorRT-LLM.
- It manages model placement, request routing, autoscaling, and weight caching.
- It is most useful for teams that want more control over and .