Show HN: SIE Serverless – managed inference powered by self-hostable stack

Category: ai-ml

Tags: inference-server, openai-compatible, self-hosted, model-serving, llm, embeddings, agents, kubernetes

Score: 7.3/10 (Innovation: 7, Technical: 8, Documentation: 7, Utility: 7)

SIE (Superlinked Inference Engine) is an open-source inference server that exposes a broad catalog of task-specific open models (embeddings, rerankers, OCR, structured extraction, classifiers, guardrails, vision, ASR, translation, agent loop) behind one OpenAI-compatible API with on-demand loading and LRU eviction. It bundles Kubernetes/Helm, KEDA autoscaling, and Grafana configs for self-hosted deployment, plus Python and TypeScript SDKs and integrations with LangChain, LlamaIndex, Haystack, DSPy, and vector DBs — a pragmatic 'one cluster for every open model your agent calls' approach.

Target audience: AI/ML engineers and backend developers deploying self-hosted agent infrastructure, plus MLOps/DevOps teams managing GPU clusters

Repository: https://superlinked.com/ · Python · Apache-2.0 · 3461 stars

View on Hacker News