San Francisco Compute is a GPU cloud that lets you sell unused compute capacity.Documentation Index
Fetch the complete documentation index at: https://docs.sfcompute.com/llms.txt
Use this file to discover all available pages before exploring further.
Development priorities
- Expanding our capabilities as a public cloud. We build core primitives for buying, managing, and orchestrating compute, then layer higher-level functionality on top.
- Increasing liquidity. If you have unused compute, you should be able to resell it and recoup some of your compute costs.
Feature states
Features move through the following stages as they mature.- In development. We are building this feature. It is not yet available.
- Private preview. Available to selected customers. We use private preview to validate designs with real usage before opening access more broadly.
- Public preview. Available for everyone to use but may have rough edges. A such, it is not covered by our uptime SLA.
- Generally available. Production-ready and fully supported under a stable version prefix. Covered by our standard SLA.
- Planned end of life. Scheduled for retirement. It still works, but you should migrate to the recommended alternative.
- End of life. No longer available or supported.
Public preview features
The following features are in public preview.- Capacities - reserve GPU capacity and manage your allocations
- Orders - place and manage compute orders
- Instances - launch and operate GPU instances
- Instance SKUs - available instance configurations
- Instance templates - reusable instance configurations
- Procurements - programmatic procurement workflows
- Deployments - deploy workloads to your instances
- Images - custom VM images
- Tags - organize resources with key-value tags
- Orderbook - view market depth and pricing
- Reselling - resell capacity you aren’t using