NeuralOpsAI is an open-source MLOps platform that streamlines the full machine-learning lifecycle, from data ingestion to training, deployment, and monitoring. It provides a unified framework for orchestrating complex ML workflows at scale.
Tagline: Operate ML like software.
- Name: NeuralOpsAI (capitalization matters)
- Short name: NeuralOps
- Tagline: Operate ML like software.
- Usage: Use the full name on first mention in docs and announcements
- Natural language pipeline creation
- Orchestration with Argo Workflows and Temporal
- Microservices architecture with gRPC and REST APIs
- Model registry with versioning and promotion stages
- AutoML for automated model selection and tuning
- Data versioning and lineage via Apache Iceberg
- Dashboard for monitoring, visualization, and health metrics
- Web Dashboard: React-based UI for pipeline creation and monitoring
- API Gateway: Entry point for REST and gRPC traffic
- Orchestrator: Parses natural language queries and compiles workflows
- AI Engine: Training, evaluation, and serving logic
- Iceberg Manager: Data lake storage, snapshots, and schema evolution
- Model Registry: Model metadata, versions, and lifecycle management
- AutoML Service: Tuning and model selection
- Incident Responder: Automated monitoring and incident actions
- Resource Optimizer: Resource allocation and cost efficiency
- Workflow Runner: Executes workflows with Argo Workflows
- Docker (for the quickest start)
- Go 1.24+ and Node.js 18+ for local builds
- kubectl and Minikube or Kind for Kubernetes development
- Buf for protobuf generation
make up
Access the dashboard at http://localhost:8080
make down
-
Install dev tools: make dev-tools
-
Generate protobufs: make proto
-
Build all services: make build
-
Run API Gateway: make run-gateway
-
Run dashboard:
cd web/dashboard npm install npm run dev
The OpenAPI spec lives at api/openapi/gateway.yaml. gRPC definitions are in api/proto.
Core endpoints:
- POST /v1/pipelines:nl
- POST /v1/pipelines/{id}:run
- GET /v1/runs/{id}
- GET /v1/pipelines
- GET /healthz
- QUICKSTART.md for setup and environment details
- ARCHITECTURE.md for system design
- IMPROVEMENTS.md for new capabilities and roadmap items
- NeuralOpsAI_Report.md for an in-depth overview
Please read CONTRIBUTING.md before submitting changes. Community standards are defined in CODE_OF_CONDUCT.md, and project decision-making is documented in GOVERNANCE.md.
See SECURITY.md for reporting vulnerabilities.
See CHANGELOG.md for release notes.
This project is licensed under the MIT License. See LICENSE for details.