aid: sf-compute
name: San Francisco Compute Company
description: San Francisco Compute Company (SF Compute) operates a market for buying short-duration time
  on H100 and H200 GPU clusters with no long-term contracts. Customers reserve VM nodes for any quantity,
  duration, and start time through the sf CLI and the sfcompute.com Orders API, with managed Slurm and bare-metal
  clusters available by request.
image: https://kinlane-productions.s3.amazonaws.com/apis-json/apis-json-logo.jpg
url: https://raw.githubusercontent.com/api-evangelist/sf-compute/refs/heads/main/apis.yml
created: '2026-05-23'
modified: '2026-05-23'
specificationVersion: '0.19'
type: Index
access: 3rd-Party
position: Producer
tags:
- AI
- Cloud
- Clusters
- Compute
- GPU
- Machine Learning
- Marketplace
- Training
apis:
- aid: sf-compute:orders-api
  name: SF Compute Orders API
  description: The SF Compute Orders API is the REST control plane for placing, listing, and managing orders
    that reserve GPU node-hours on H100 and H200 clusters. Orders specify node count, duration, and start
    time. Authentication uses bearer tokens generated via the sf CLI.
  humanURL: https://docs.sfcompute.com
  baseURL: https://api.sfcompute.com/v0
  tags:
  - Clusters
  - GPU
  - H100
  - H200
  - Orders
  - REST
  properties:
  - type: Documentation
    url: https://docs.sfcompute.com
common:
- type: Website
  url: https://sfcompute.com
- type: Developer
  url: https://docs.sfcompute.com
- type: Documentation
  url: https://docs.sfcompute.com
- type: SignUp
  url: https://sfcompute.com/signup
- type: Login
  url: https://sfcompute.com/signin
- type: Pricing
  url: https://sfcompute.com/prices
- type: Blog
  url: https://sfcompute.com/blog
- type: ChangeLog
  url: https://sfcompute.com/changelog
- type: Support
  url: https://sfcompute.com/contact
- type: GitHubOrganization
  url: https://github.com/sfcompute
- type: LinkedIn
  url: https://www.linkedin.com/company/sf-compute
- name: sf CLI
  url: https://github.com/sfcompute/cli
  type: CLI
- name: Nodes API Go SDK
  url: https://github.com/sfcompute/nodes-go
  type: SDK
- name: Nodes API TypeScript SDK
  url: https://github.com/sfcompute/nodes-typescript
  type: SDK
- name: SFC Go SDK
  url: https://github.com/sfcompute/sfc-go
  type: SDK
- name: SFC TypeScript SDK
  url: https://github.com/sfcompute/sfc-sdk-typescript
  type: SDK
- type: Features
  data:
  - name: Market-Based GPU Pricing
    description: Per-node hourly pricing fluctuates with supply and demand, with no long-term contracts.
  - name: Flexible Node Reservations
    description: Reserve any number of H100 or H200 nodes for custom durations and start times.
  - name: Managed Slurm Clusters
    description: Slurm-scheduled clusters available on request for larger training jobs.
  - name: Bare Metal Clusters
    description: Direct bare-metal access for advanced workloads on request.
  - name: Hardware Failure Refunds
    description: Automatic refunds for failed nodes with hardware-level monitoring.
  - name: Cancel Anytime
    description: Cancel or resell unused capacity without penalty.
- type: GPUs
  data:
  - name: NVIDIA H100
    description: 80GB Hopper GPU nodes available with market-based pricing (around $1.97/hr average).
  - name: NVIDIA H200
    description: 141GB HBM3e Hopper GPU nodes available for hourly and longer-term rental.
  - name: NVIDIA B300
    description: Blackwell Ultra GPU nodes listed as upcoming on the SF Compute roadmap.
- type: Customers
  data:
  - name: NVIDIA
  - name: Roboflow
  - name: Datology
  - name: Liquid AI
  - name: MIT
- type: LLMsTxt
  url: https://docs.sfcompute.com/llms.txt
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com
  url: https://apievangelist.com