# vllm-project/aibrix

Cost-efficient and pluggable Infrastructure components for GenAI inference

Repository: https://github.com/vllm-project/aibrix
Canonical: https://ross.abutalabs.com/products/aibrix
Language: Go
License: Apache-2.0
License Family: permissive
Last push: 2026-08-26T19:43:52+00:00

## Health v2 (maintenance only)
Score: 83/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 77, longevity 58
- inputs: {"age_days": 814, "days_push": 7, "days_rel": 77, "gap_med": 60.5, "n_releases_24m": 11}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5038, forks 663 (observed 2026-08-28T04:09:08.756353+00:00)

## What it is
AIBrix is an open-source, cloud-native framework from the vLLM project providing essential building blocks for scalable, cost-efficient GenAI/LLM inference infrastructure on Kubernetes. It offers a pluggable control plane with components for model serving, autoscaling, routing, and gateway management tailored to enterprise LLM deployments.

## Use cases
- deploy and scale vLLM LLM inference on Kubernetes
- build a cost-efficient LLM serving control plane
- autoscale large language model inference workloads
- route traffic across multiple LLM model replicas
- run DeepSeek-R1 or other large models in production
- manage GenAI inference infrastructure for enterprise

## When to choose
- you serve LLMs at scale on Kubernetes and need autoscaling, routing, and gateway features
- you want a Kubernetes-native control plane purpose-built for vLLM inference
- you need cost-efficient multi-model LLM serving infrastructure

## When to avoid
- you only need to run a single LLM locally without Kubernetes
- you use a managed LLM API and don't self-host models
- your stack is not container/Kubernetes based

## Facets
- artifact type: framework
- maturity: active
- function: llm-inference, container-orchestration, api-gateway, deployment, monitoring, routing
- domain: large-language-models, infrastructure-as-code, cloud-computing, microservices
- platform: go, cloud, self-hosted
- tags: genai-inference, vllm, llm-serving, control-plane, autoscaling, model-serving, kubernetes-operators, devops, kubernetes, docker

## Member repositories
- vllm-project/aibrix (main) score 83

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:08.756353+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 2026-08-29T18:17:25.288691+00:00, confidence not recorded.
  - readme: https://github.com/vllm-project/aibrix (fetched 2026-08-28T04:09:08.756353+00:00, sha 6933c708ea57)
- Data as of 2026-08-30T08:39:29.467469+00:00.