# Runcrate

Source: [Runcrate](<https://www.sota2.com/companies/runcrate>)

Collection index: [llms.txt](<https://www.sota2.com/companies/llms.txt>)

Data updated: 2026-11-17T23:38:50.452Z

On-demand GPUs and inference for AI teams and the agents they run, including GPU machines with real shells and an OpenAI-compatible model API.



## Company facts

| Field | Recorded value |
| --- | --- |
| Local name | Not recorded |
| Categories | Models & AI Infrastructure; Software Development |
| Headquarters | Not recorded |
| Deployment | Not recorded |

## Products by this company

- [GPU Compute](<https://www.sota2.com/products/runcrate-gpu-compute.md>): Named-GPU cloud instances and crates with a real shell for training, fine-tuning, serving, research, and agent jobs; billed for runtime.

- [Inference API](<https://www.sota2.com/products/runcrate-inference-api.md>): An OpenAI-compatible API for hosted open and frontier models, including language, image, video, speech, and audio models, billed per use.

- [Storage Volumes](<https://www.sota2.com/products/runcrate-storage-volumes.md>): Persistent volumes that attach to GPU instances across regions, keep files after instance termination, and provide local-disk-speed access.

- [Reserved GPU Clusters](<https://www.sota2.com/products/runcrate-reserved-gpu-clusters.md>): A cluster reservation service that matches multi-node GPU requirements with GPU-owner offers and delivers the selected cluster into the Runcrate dashboard.

## Task-specific ranking appearances

## Original provider information

[Official website](<https://www.runcrate.ai/>)
