Seldon is now
TrueFoundry

The world's most battle-tested Kubernetes MLOps stack now combines with TrueFoundry's enterprise AI platform — giving you a single, cloud-agnostic foundation from real-time inference to Agentic AI.

Read the Announcement →

Trusted by the world's most innovative teams building real-time ML and AI · 25,000+ MLOps professionals

iris-pipeline.yaml — live

Inference pipeline • running

model_request
→
feature_transform
iris_classifier
→
prediction
outlier_detector
→
alibi_explain
2M+
Installs
40+
Backends
10yr
Production

Better together

Kubernetes-first, cloud-agnostic, developer-native — unlocking the next generation of enterprise AI.

Read technical documentation →
🤖

Agentic AI

Move beyond static ML models. Deploy autonomous AI agents that reason, plan, and act in enterprise workflows — without ripping out your Kubernetes infrastructure.

☸️

Kubernetes-native, everywhere

A shared architectural north star. Portable, scalable, and secure deployment across any cloud or on-premise environment. No vendor lock-in — ever.

🚀

Accelerated path forward

Existing Seldon enterprise customers get a clear, low-risk path to Agentic and Composite AI. TrueFoundry reduced enterprise deployment timelines by 50%+ for global F500 firms.

Top 4 Open-Source AI Deployment Tool 2026Independent ranking — Seldon

Seldon is trusted by the world's most innovative teams building real-time machine learning and AI

Capital One Covea AstraZeneca GSK Aselsan Noda Cambridge University
Products

The complete MLOps toolkit

From open-source inference serving to enterprise governance — Seldon covers the full production ML lifecycle. Ranked #4 best open-source AI deployment tool of 2026.

Seldon Core 2
Open Source

A modular, data-centric framework for deploying and scaling ML and LLMOps in production. Kubernetes-native with real-time Kafka data pipelines.

Explore Core 2
MLServer
Open Source

A lightweight inference server that receives input data, runs it through trained models, and returns predictions. Fully compatible with the Open Inference Protocol.

Explore MLServer
LLM Module
Module

Deploy GenAI workflows with observability, prompt orchestration, and production-ready scaling. Built-in and configurable guardrails for responsible LLM deployment.

Explore LLM Module
Enterprise Platform
Enterprise

Comprehensive oversight and governance for ML and LLM deployments. Enhanced authentication, audit trails, and team controls for regulated industries.

Explore Enterprise
Model Performance Metrics
Module

Optimize production classification and regression models with real-time quality insights. Detect performance degradation before it impacts business outcomes.

Explore MPM
Alibi Detect
Alibi Explain
Open Source

Outlier, adversarial, and drift detectors alongside global, local, black-box, and white-box explanation methods. Trusted by regulated industries worldwide.

Explore Alibi
Platform capabilities

Built for the realities of production ML

Seldon Core 2 is designed around the operational challenges AI Platform Engineering teams face once models leave the notebook — real-time inference, observability, and compliance at scale.

🔧

Model deployment at any scale

Orchestrate ML components as Kubernetes microservices. Deploy locally or at enterprise scale with REST and gRPC support for every model type.

🔗

Composable data-centric pipelines

Connect models, processing steps, custom logic, and monitoring components into real-time Kafka-powered pipelines. Every byte of data is observable.

🧪

A/B tests and canary deployments

Run routing experiments in production — A/B tests, canary deployments, multi-armed bandits. Promote winners without downtime.

📈

Real-time observability

Every prediction logged and auditable. Monitor data pipelines, model performance, and deployments with Prometheus, Grafana, and custom dashboards.

💰

Multi-model serving and overcommit

Consolidate models onto shared inference servers. LRU memory swapping provisions more models than hardware allows — cutting GPU costs significantly.

☁️

Cloud-agnostic, on-premise ready

Tested on AWS EKS, Azure AKS, Google GKE, Alicloud, Digital Ocean, and OpenShift. Your data stays wherever you need it.

Deploy in minutes

Standardized inference.
Any model type.

One Kubernetes manifest. Automatic inference server selection, scaling, monitoring, and audit logging — all built in.

View install guide
sklearn-iris-pipeline.yaml 
# Deploy a Scikit-learn model
apiVersion: mlops.seldon.io/v1alpha1
kind: Model
metadata:
  name: iris
spec:
  storageUri: "gs://seldon-models/scv2/iris-sklearn"
  requirements: [sklearn]
  memory: 100Ki

---
# Wire it into an observable pipeline
apiVersion: mlops.seldon.io/v1alpha1
kind: Pipeline
metadata:
  name: iris-pipeline
spec:
  steps:
    - name: iris
    - name: outlier-detector
  output:
    steps: [iris]
Why Seldon

Production AI at enterprise scale

Purpose-built for VP Engineering, Chief AI Officers, and AI Platform Architects who need more than a model server — governance, explainability, and zero vendor lock-in.

🔓

Avoid Lock-In

Deploy and scale any AI model across any cloud or on-premise. Never limited by vendor lock-in or integration gaps.

🚀

From POC to Production

Seldon Core 2 standardizes complex AI deployments with Kubernetes-native pipelines, making ML and GenAI production-ready out of the box.

🛡️

Trust Every AI Decision

Real-time monitoring, drift detection, and explainability so your AI systems stay transparent, reliable, and compliant at scale.

💰

Cut Costs, Not Performance

Multi-model serving and overcommit consolidate workloads to slash infrastructure cost while keeping latency low and throughput high.

🧪

Model Experimentation

A/B testing to shadow deployments — simple, scalable, and disruption-free for seamless real-time machine learning.

🧩

Modular Architecture

Budget accurately and only pay for what you need. Seldon's modular design extends to pricing — from open-source to enterprise.

Integrations

Fits into your existing stack

Seldon integrates with the tools your team already uses — from CI/CD to monitoring, storage to service mesh.

PrometheusGrafanaKafka JaegerElasticsearchTriton Inference MLflowWeights & BiasesIstio EnvoyArgoCDFlux AWS EKSAzure AKSGoogle GKE OpenShiftHugging FaceLangSmith rclone (40+ backends)Open Inference Protocol DatabricksDigital Ocean
Documentation

All Seldon products, one place

The full ecosystem is documented at docs.seldon.ai — from quickstart to advanced production patterns.

ProductDescriptionTypeDocs
Seldon Core 2Kubernetes-native MLOps and LLMOps framework — the core deployment engine.Open Sourcedocs →
MLServerLightweight multi-framework inference server with REST/gRPC and Open Inference Protocol.Open Sourcedocs →
LLM ModuleGenAI deployment with prompt orchestration, observability, and guardrails.Moduledocs →
Alibi DetectOutlier, adversarial, and drift detection — broad coverage of model and data types.Open Sourcedocs →
Alibi ExplainModel inspection and explanation — local, global, black-box, and white-box methods.Open Sourcedocs →
MPM ModuleModel Performance Metrics — production quality insights for classification and regression.Moduledocs →
Enterprise PlatformOversight, governance, enhanced authentication, and audit for ML and LLM at scale.Enterprisedocs →
Core 1 (legacy)Cloud-agnostic open-source platform for deploying ML models — the original Seldon Core.Legacydocs →
Why teams choose Seldon

Finally get real-time AI into production

Unlock new opportunities for innovation and growth — without ripping out your existing infrastructure.

"With our Model as a Service platform running on Seldon, we've gone from it taking months to minutes to deploy or update models."

Enterprise Customer

"Seldon has made a huge difference to how we scale and deploy our inference ecosystem."

Enterprise Customer

"Seldon enables us to productionize models at speed while adding explainers into every one — pivotal to our mission of becoming the most advanced AI Factory in the industry."

Enterprise Customer
What's next

Your Kubernetes stack is ready
for what comes next

Existing Seldon enterprise customers get a direct, low-risk path to Agentic and Composite AI — on the same infrastructure you already run in production. Schedule a briefing to see the combined TrueFoundry + Seldon roadmap.

Schedule a platform briefing Read the Announcement →
TrueFoundry

TrueFoundry + Seldon is the most complete enterprise AI deployment platform on Kubernetes — from real-time MLOps to Agentic AI, with no vendor lock-in and full cloud portability. Proven at NVIDIA, Fortune 500, and global enterprises. Explore the combined platform →