#6969th of 71 in Model serving
seldon-core
Kubernetes framework for serving ML models and Kafka-connected pipelines
- Stars
- 4.8k
- License
- custom license
- Last commit
- Jan 2026
- Last release
- Dec 2025
- Language
- Go
4,726 on 22 Apr 20264,785 on 11 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Seldon Core 2 deploys and scales models on Kubernetes, with multi-model serving on shared inference servers and overcommit to host more models than fit in memory. Pipelines chain models and custom components using Kafka for streaming data, and experiments route traffic for A/B tests and shadow deployments. It is written in Go and distributed under the Business Source License.
Who it is for: Platform and MLOps teams serving many models on Kubernetes
Strengths
- Multi-model serving consolidates several models on shared inference servers
- Overcommit lets you deploy more models than memory allows
- Experiments support A/B tests and shadow deployments between models or pipelines
- Pipelines stream data between components over Kafka
Weaknesses
- Business Source License, not OSI open source; commercial use may require Seldon
- Requires a Kubernetes cluster; no standalone or Docker quick start in the README
- Pipelines depend on Kafka, adding another system to operate
- No hardware requirements or model format list in the README
What it needs
- Needs Kubernetes, Kafka
What changed
CVE-2025-61729. “Multiple CVEs addressed across Go, Python, and container images, including the operator, executor, alibi-detect server, python-builder, and core builder images.”
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in Model serving
See all 71| Rank | Project | Score |
|---|---|---|
| 1 | llama.cppC/C++ inference engine for GGUF LLMs and VLMs with an OpenAI-compatible server | 94 out of 100 |
| 2 | OllamaLocal model runner with CLI and REST API for open-weight LLMs | 92 out of 100 |
| 3 | LocalAISelf-hosted inference server with OpenAI-compatible API and pluggable backends | 91 out of 100 |
| #4 | vLLMInference engine serving LLMs with PagedAttention and an OpenAI-compatible API | 88 out of 100 |
| #5 | strataRuns a 125B-parameter MoE model on one consumer GPU plus system RAM | 88 out of 100 |