zgiai/zgi

ZGI is an open-source platform for building AI applications. Its intuitive interface combines workflow design, agent orchestration, dataset management, and model integration—allowing you to quickly move from prototype to production.

What it solves

ZGI is a self-hostable Agent Runtime platform designed for teams that need AI agents to perform actual work rather than just chatting. It provides a unified workspace to build, connect, and operate AI agents and executable workflows, bridging the gap between a simple LLM prompt and a production-ready application with governance and observability.

How it works

ZGI allows users to configure agents with specific instructions, memory, and knowledge bases. It uses a visual canvas for orchestrating executable workflows that include LLM calls, database operations, HTTP requests, and human-in-the-loop approvals. The platform includes a model gateway for routing and quota management, a sandboxed runtime for executing tools and code, and a retrieval system that combines semantic, full-text, and knowledge-graph search.

Who it’s for

It is intended for teams and organizations that want to deploy self-hosted AI agents that are grounded in company files and business data, and who require control over permissions, runtime logs, and execution governance.

Highlights

  • Visual Workflow Orchestration: A drag-and-drop canvas to build complex agentic processes including loops and branches.
  • Advanced Retrieval: Hybrid search capabilities (semantic, full-text, and knowledge-graph) with reranking to ensure grounded responses.
  • Model Gateway: Centralized management of LLM providers, routing policies, and pricing metadata.
  • Sandboxed Execution: Isolated runtime for executing reusable skills, calculations, and code.
  • Self-Hosted Infrastructure: Full stack deployment using Docker Compose, including PostgreSQL and Redis.

Related

  • Project
    EpicStaff/EpicStaff

    EpicStaff is a self‑hosted, source‑available platform that lets teams design, run, and audit multi‑agent AI workflows via a visual node editor built on Django. It supports custom Python tools, LLM/RAG integration, persistent memory, webhook triggers, and enterprise security, giving operations teams ownership of AI processes while engineers handle system integration.

  • Project
    Agenta-AI/agenta

    Agenta is an open‑source, self‑hostable workspace for building AI agents that can be chatted with, shared with a team, and run in the background. It supports many LLM providers, integrates with thousands of SaaS tools, offers permission controls, tracing, and scheduling, and provides a web UI plus a rich developer ecosystem.

  • Dispatch
    Runtime: Scaling Agentic Workflows Across the Enterprise

    An exploration of Runtime, a sandboxed infrastructure for deploying coding agents across non-engineering teams while maintaining strict security and governance.

  • Project
    truefoundry/trueforge

    TrueForge is an open‑source TypeScript runtime that executes LLM agents. It manages model calls, tool invocations, sandboxed code, approvals, and session state, exposing a chat UI, HTTP API, and TypeScript SDK. Deploy locally (SQLite) or in hosted mode (Postgres + Redis) via Docker‑Compose, Helm, or Railway.

  • Project
    erthorpabar/pabar_agent_runtime

    An AI agent runtime framework that integrates chat, RAG, and agentic loops to enable professional automation tasks across desktop and mobile platforms.