OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Infrastructure & Serving

Superlinked Inference Engine

Superlinked Inference Engine (SIE) is an open-source inference server and cluster for hosting open models used by agent applications. It provides an OpenAI-compatible API and serves tasks including search and retrieval, document-to-markdown conversion, structured output, classification, content checks, translation, vision and agent-loop generation. Its catalog contains more than 100 models, with on-demand loading and

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Superlinked Inference Engine — OpenViST project profile in AI Infrastructure & Serving.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use SIE when an agent application needs open models served from a self-hosted environment and a shared inference interface for multiple model-backed tasks.

Who it is for

  • Teams building agent applications that need self-hosted inference for open models.
  • Engineers consolidating multiple model-backed tasks behind an inference API.
  • Operators deploying model serving locally or in cloud infrastructure, including Kubernetes-based clusters.

Documented uses

  • Serve embeddings and text-generation requests through an OpenAI-compatible API.
  • Run model-backed retrieval, document conversion, structured extraction, classification, translation and vision tasks.
  • Host the open models used by an agent's planning and tool-calling loop.
  • Deploy and scale inference services using the supplied Kubernetes and Helm configurations.
Before you choose

Requirements & limits

The project is open source under the Apache-2.0 license. The quickstart documents local Python 3.12 setup on macOS with Apple Silicon or Linux, plus Docker options for Linux CPU and NVIDIA GPU.

  • The quickstart's native setup requires Python 3.12 and is documented for macOS with Apple Silicon or Linux.
  • Docker images differ by dependency bundle; some OCR and decision models require the Transformers 5 or SGLang vision images.
  • The first request for each model downloads its weights.
  • The API is OpenAI-compatible, including embeddings and generation endpoints.
  • The repository lists Kubernetes and Helm deployment configurations for gateway load balancing, KEDA autoscaling and Grafana dashboards.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Start SIE on Linux with an NVIDIA GPU

Launch the documented default CUDA server image and confirm that the local server is ready.

Before you start

  • Linux host with an NVIDIA GPU
  • Docker with GPU access configured

Steps

  1. Run the documented Docker command using the latest-cuda12-default image, map port 8080 and mount the sie-hf-cache volume.

  2. In a second terminal, request http://localhost:8080/readyz with curl and check for the documented ok response.

Expected outcome

A running SIE server at port 8080 with readiness checked.

Guide limitations

  • This setup uses the default CUDA image; the README specifies separate images for certain OCR and model profiles.
  • This guide does not cover cluster deployment or agent integration.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
10 October 2026
Original website title
superlinked/sie · Official homepage title · retained in its source language
How we review sourcesSuggest a correction