Maximize Cloud Cost Efficiency and Performance for AI & Data Workflows.
Open Data
Python
Benchmarks
Add-on Services
PRICE/HOURS
CPU
RAM (GB)
<
extract_transform_load
>
We closely monitor cloud and flexible server providers, tracking details such as their data centers, availability zones, server offerings, and in-depth pricing options for servers, storage, traffic, IPv4 addresses etc. This comprehensive and standardized dataset is openly accessible through our public homepage, databases, APIs, and SDKs.
Here are some key highlights regarding the information we currently cover.
SQL query to show statistics on the Spare Cores inventory and update frequency
5 (out of 10 planned) vendors
353 availability zones
2344 server types
970,000+ benchmark scores
~310,000 live price records
~5k records updated hourly
~45M historical records
Tracked every 5 mins.
<
resource_optimization
>
Only 3 Steps to optimize your compute resources
Screenshot of the server listing page, focusing on the filter elements in the sidebar
1
Step
Define Your Needs
Utilize our platform's dozens of filters to precisely outline your workload requirements, specifying details like number of vCPUs, amount of memory/storage, or GPU options.
Screenshot of the server listing page, focusing on the search results, showing 3 instance records in a table
2
Step
Find Optimal Resource
Let Spare Cores narrow down 2,000+ server types and 300,000 pricing options based on your criteria. We prioritize cost efficiency, presenting you with optimal choices.
Running the SC runner CLI tool to start a t4g.nano instance
3
Step
Deploy with Ease
Use our CLI tool to launch the selected instance in your cloud environment. Our streamlined process ensures it is up and running without unnecessary complexities.

Besides our open-source projects, we are also working on a Container-as-a-Service platform to manage the full lifecycle of your containers — including provisioning, monitoring, scaling and scheduling without direct vendor engagement or any infrastructure operations. This is ideal if you do not have the capacity to manage your own cloud environment, but want to leverage the benefits of cloud computing in a cost-efficient and seamless way.

<spare_cores> is 100% committed to using and contributing to open-source!
Spare Cores is the successor to cloudperf, an AWS-focused open dataset and toolset for exploring spot prices and stress-ng benchmarks, which we started in 2017. This next-generation project was launched in Q3 of 2023 and gained momentum in 2024, particularly due to the support from NGI Search funding.
Banner image showing NGI (Next Generation Internet) funding for the SC (Spare Cores) project
<
project_components
>
The status and brief descriptions of all the existing open-source Spare Cores components:
COMPONENTSTATUSREPOSITORYDESCRIPTION
SC CrawlerBetaPython logoInventory cloud resources into a SQlite database.
SC InspectorBetaInspect and benchmark cloud resources.
SC DataBetaPython logoWrapper around data collected using the Crawler.
SC KeeperAlphaAPI logoAPI to search the Data.
SC ScannerStableWWW logoWeb frontend and programming language SDKs for Keeper.
SC RunnerBetaLaunching actual cloud instances.
<
fun_facts
>
The most expensive server costs over $250 per hour. It comes with 448 vCPUs and 24TB RAM.
Most regions have 3 independent zones.
Some are single-zone, and one have 6 zones.
While spot prices are relatively stable nowadays, they can still fluctuate by up to 10% within a day.
There is a 2500x performance difference between the fastest and slowest nodes.
Some vendors charge 6+ times the price of a public IPv4 address compared to others.
<
release_notes
>
We publish case studies, learnings from implementing Spare Cores components, and other project updates from time to time. Find some of our featured articles below:
  • ✅ Amazon Web Services (Done)
  • ✅ Google Cloud Platform (Done)
  • ✅ Hetzner Cloud (Done)
  • ✅ Microsoft Azure (Done)
  • ✅ UpCloud (Done)
  • ⚙️ Vultr (In Progress)
  • 🗓️ Oracle Cloud Infrastructure (Planned)
  • 🗓️ Scaleway (Planned)
  • 🗓️ Alibaba Cloud (Planned)
  • 🗓️ OVH Cloud (Planned)