# NVIDIA/aistore

AIStore: scalable storage for AI applications

Repository: https://github.com/NVIDIA/aistore
Canonical: https://ross.abutalabs.com/products/aistore
Homepage: https://docs.nvidia.com/aistore
Language: Go
License: MIT
License Family: permissive
Topics: object-storage, etl-offload, linear-scalability, batch-jobs, kubernetes, distributed-storage, high-availability, high-performance, ml-training, multi-cloud, multipart-upload, s3-compatible
Last push: 2026-08-26T20:41:43+00:00

## Health v2 (maintenance only)
Score: 95/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 86, longevity 100
- inputs: {"age_days": 3185, "days_push": 7, "days_rel": 12, "gap_med": 31.0, "n_releases_24m": 19}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1915, forks 283 (observed 2026-08-28T04:05:53.694572+00:00)

## What it is
AIStore (AIS) is a lightweight, distributed object storage stack built by NVIDIA specifically for AI workloads. It provides linear scalability, multi-cloud access (S3, GCS, Azure, OCI), S3-compatible API, ETL capabilities, and can be deployed anywhere from a single Linux machine to petascale Kubernetes clusters.

## Use cases
- store training datasets for ML models at scale
- serve S3-compatible storage for AI pipelines
- accelerate multi-cloud data access for deep learning
- run ETL transformations on stored objects
- deploy scalable storage on Kubernetes for AI clusters
- batch download large datasets for model training

## When to choose
- you need high-performance storage optimized for ML training workloads
- you want a single namespace across multiple cloud providers
- you need linear scalability with erasure coding and mirroring
- you want built-in ETL and batch object retrieval

## When to avoid
- you need a simple single-node file server
- your workload is small and a plain S3 bucket suffices
- you require POSIX filesystem semantics rather than object storage

## Facets
- artifact type: service
- maturity: active
- function: object-storage, etl, caching, monitoring, benchmarking
- domain: machine-learning, databases, cloud-computing, big-data, self-hosted, microservices
- platform: go, python, self-hosted
- tags: s3-compatible, multi-cloud, ai-workloads, high-performance-storage, erasure-coding, pytorch-integration, linux, docker, kubernetes

## Member repositories
- NVIDIA/aistore (main) score 95

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:53.694572+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:10:18.497308+00:00, confidence not recorded.
  - readme: https://github.com/NVIDIA/aistore (fetched 2026-08-28T04:05:53.694572+00:00, sha db7182fd6c2d)
  - homepage: https://docs.nvidia.com/aistore (fetched 2026-08-29T10:49:38.526247+00:00, sha f0d266dbef92)
  - site_page: https://docs.nvidia.com/aistore/docs (fetched 2026-08-29T10:49:38.538546+00:00, sha ac27c2d83f26)
  - site_page: https://docs.nvidia.com/aistore/cli/etl (fetched 2026-08-29T10:49:38.540978+00:00, sha c74e33127461)
  - site_page: https://aistore.nvidia.com/docs/http-api (fetched 2026-08-29T10:49:38.543422+00:00, sha 48a3bb20e3b4)
  - site_page: https://www.nvidia.com/en-us/about-nvidia/privacy-policy (fetched 2026-08-29T10:49:38.545699+00:00, sha 5362c58d0750)
  - site_page: https://www.nvidia.com/en-us/about-nvidia/privacy-center (fetched 2026-08-29T10:49:38.549679+00:00, sha b098377da9cc)
  - site_page: https://www.nvidia.com/en-us/about-nvidia/terms-of-service (fetched 2026-08-29T10:49:38.551973+00:00, sha 85469b6ff0a1)
  - site_page: https://www.nvidia.com/en-us/about-nvidia/accessibility (fetched 2026-08-29T10:49:38.554491+00:00, sha 8f61ce4143ec)
  - site_page: https://www.nvidia.com/en-us/about-nvidia/company-policies (fetched 2026-08-29T10:49:38.556109+00:00, sha b169611fd7d6)
- Data as of 2026-08-30T08:39:29.467469+00:00.