DataCrunch
DataCrunch is a European (Finland-based) GPU cloud offering on-demand and reserved NVIDIA GPU instances (H200, H100, B200, A100, L40S, V100) plus a serverless inference and container deployment platform. Its REST API at https://api.datacrunch.io/v1 uses OAuth2 client-credentials to issue Bearer tokens and exposes instances, instance types, availability, images, SSH keys, startup scripts, volumes, balance, and serverless containers, with an OpenAI-compatible inference endpoint for deployed models.
APIs
DataCrunch Instances API
Deploy, list, inspect, and act on GPU/CPU instances. Actions (boot, shutdown, hibernate, restore, delete, force-shutdown) are issued via PUT against one or more instance IDs.
DataCrunch Instance Types & Availability API
Lists available instance types with GPU/CPU/RAM/storage specs and per-hour pricing, and reports real-time availability across datacenter locations.
DataCrunch Images & Startup Scripts API
Lists available OS images for instances and clusters, and manages startup (cloud-init) scripts used to provision instances at deploy time.
DataCrunch SSH Keys API
Add, list, retrieve, and delete the SSH public keys injected into instances at deploy time for secure access.
DataCrunch Volumes API
Create, list, inspect, and delete block storage volumes and run volume actions (attach, detach, resize, clone, rename), including a trash listing for soft-deleted volumes and a ...
DataCrunch Serverless Inference & Containers API
Deploy and manage autoscaling serverless container deployments (create, list, get, update, delete, scaling, status, pause, resume, purge-queue) running vLLM/TGI/SGLang, and invo...
DataCrunch Balance API
Retrieves the project account balance and currency for the authenticated client.