Skip to main content
Task-Oriented Docs

What would you like to achieve with OpenCSG today?

Search technical solutions, explore hands-on tasks, or start from a role-tailored pathway.

⌘ K
Deploy Enterprise AIPrivate cloud setup, architecture design & governance
Build Your First AgentLaunch an agent instance from scratch or templates
Manage Models & DatasetsUnified hosting, Git LFS transfer & version audit

Start by Goal & Persona

Choose the learning pathway tailored to your team scale and operational needs.

All Pathways

Enterprise LLM & Agent Platform Infrastructure

Private Cloud & Governance

For enterprise IT and platform engineering teams deploying sovereign, compliant, and highly available AI infrastructure.

Recommended Path & Milestones
01
Architecture & Hardware SizingCompute, storage sizing & domestic NPU compatibility
02
Airgap Offline DeploymentOffline bundle import & MinIO/Ceph object storage
03
Multi-Tenant Permissions & IsolationWorkspaces, member roles & security boundaries
04
Production Support & DiagnosticsSystem health monitoring, log collection & recovery
Request Enterprise Evaluation View Complete Pathway

Core Scenarios & Ecosystem Channels

Key capabilities built for enterprise sovereignty, airgap environments & developer ecosystems.

Core Advantage

Domestic Compute & NPU Support

Optimized for Huawei Ascend NPU, Enflame, and Hygon DCU with drivers, container images, and quantized inference.

Explore Guide
Production Ready

Airgap & Isolated Enterprise Setup

For high-security financial and government environments, supporting offline bundle imports, MinIO/Ceph, and network isolation.

Explore Guide
Open Standard

Model Context Protocol (MCP)

Connect AI agents securely with internal enterprise databases, business systems, and APIs via standard MCP.

Explore Guide
Asset Hub

Open Model & Dataset Hub

Host open-source foundational models and datasets with Git LFS acceleration, data cleaning, and online spaces.

Explore Guide

Hands-on Practical Tasks

Start from real objectives and deliver working, verifiable outcomes in set timeframes.

Explore All Tasks
01
Deploy CSGHub On-PremiseQuickly launch enterprise-grade model asset management on local servers or private clouds.
30–60 minLevel: IntermediateProduct: CSGHub
02
Create & Publish First AgentConfigure prompts, models, and built-in tools through simple UI to publish a functional agent.
15–20 minLevel: BeginnerProduct: AgenticHub
03
Enterprise RAG & Knowledge BaseParse multi-format documents, vectorize content, bind agents, and verify hybrid retrieval accuracy.
30–45 minLevel: IntermediateProduct: AgenticHub
04
Configure Multi-Tenant PermissionsEstablish team roles, resource isolation boundaries, and least-privilege security policies.
20–30 minLevel: BeginnerProduct: CSGHub
05
Deploy Scalable Model InferenceAllocate GPU compute, spin up inference endpoints, and test OpenAI-compatible REST APIs.
30–40 minLevel: IntermediateProduct: API & SDK
06
Manage LLM Weights & SnapshotsHost large model weights via Git LFS with version tracking, integrity audit, and resumable transfers.
15–20 minLevel: BeginnerProduct: CSGHub

Close the Production Loop with AgenticOps

From business intent to continuous fine-tuning: standardized agentic lifecycle methodology.

Explore AgenticOps Methodology
  1. 01PromptIntent & Requirements Planning
  2. 02CodeAssisted Coding & API Calls
  3. 03BuildAssemble Agent Workflows
  4. 04TestBehavioral Benchmark & Verification
  5. 05ReleaseArtifact Packaging & Publishing
  6. 06DeployDeploy to Production / Sovereign Infra
  7. 07OperateObservability & Token Operations
  8. 08RetrainData Feedback & Model Optimization

Complex Deployments & Governance — Plan with OpenCSG Experts

We provide high-availability architecture advisory, sovereign compute compatibility, and on-site POC support.

Get Enterprise Deployment Evaluation