Discover and explore top open-source AI tools and projects—updated daily.
William-Lu-stackAI-native SRE control plane for Kubernetes and cloud infrastructure
New!
Top 41.3% on SourcePulse
Summary
LuxyAI is an AI-native SRE control plane for Kubernetes and cloud infrastructure, enabling systems to self-explain, safely self-heal, and audibly prove recovery. It targets SREs and operations teams, offering an auditable "AgenticOps Loop" integrating alerts, evidence, topology, human approval, and controlled remediation to enhance infrastructure reliability.
How It Works
The core "AgenticOps Loop" emphasizes "Evidence First," "Guarded Action," and "Verified Recovery." It ingests diverse data (alerts, logs, metrics, topology) for diagnosis. Actions are strictly controlled via RBAC, policies, dry-runs, and human approval. Recovery is validated by re-testing the original symptom, and the system learns from failed remediation strategies, persisting lineage across restarts.
Quick Start & Requirements
Installation involves cloning, configuring LLM API endpoints, building the frontend, and running via uvicorn. Docker and Kubernetes (Helm) deployments are supported. Prerequisites include Python, Node.js, and an LLM gateway. Local setup requires configuring .env. Links to the GitHub repo and "Field Notes" are available.
Highlighted Details
Maintenance & Community
Created by 陆宣宇 (Xuanyu Lu), LuxyAI is presented as a project "building in public" with a roadmap to evolve into an AgenticOps operating system. A bilingual "Field Notes" series details its development. Specific community channels are not listed.
Licensing & Compatibility
Released under the PolyForm Noncommercial License 1.0.0. Permits non-commercial use, study, modification, and redistribution. Commercial use, hosted services, resale, or product bundling require explicit written authorization from the maintainer. Not an OSI-approved open-source license.
Limitations & Caveats
The primary limitation is the non-commercial license, restricting adoption in commercial products without separate agreement. High-risk actions require explicit operator confirmation. Custom algorithm extensions, credentials, and production topology data are excluded from the public repository.
1 day ago
Inactive