Projects

LightStruc AI

The latest project our team at GVN has been developng is LightStruc AI (LSAI). It's a full-stack AI orchestration platform designed to simplify the building, deployment, and management of multi-agent AI workflows by providing a visual environment to connect AI agents, external tools, and large language models without requiring extensive boilerplate code. Chris Daugherty, CTO at GVN, expressed a need to find a better way to connect up our customer pipeline requirements while syncronizing and interpreting inter-agent chatter. After a few months of spitballing ideas we began the task of integrating our existing agentic libraries, templates, tools and vendor APIs with the new web based GUI that now hosts our tool-chains. Additional information is available at lightstruc.com.

LSAI STUDIO · VISUAL WORKFLOW BUILDER

Core Features & Functionality

Visual Workflow Builder

A drag-and-drop canvas that allows users to design multi-agent pipelines, wire components together, and configure system prompts visually. It features color-coded nodes, live graph previews, and auto-layout tools.

Multi-Agent Orchestration

Enables the chaining of specialized AI agents to handle complex tasks. The platform automatically manages context propagation, routing between agents, and error recovery.

Model-Agnostic Routing

Supports integration with over 10 AI providers, including OpenAI, Anthropic, Google, AWS Bedrock, Ollama, and vLLM. Users can switch models on a per-agent basis and implement intelligent load balancing and failover routing.

Tool & MCP Integration

Agents can be connected to external capabilities using the Model Context Protocol (MCP), REST APIs, and webhooks. The platform automatically discovers available tools and allows one-click enabling.

Secure Relay Tunneling

Provides encrypted tunnels to securely connect cloud workflows to private network resources — such as local Ollama instances or internal enterprise databases — without requiring VPNs or firewall modifications.

Real-Time Observability

Built-in analytics track token usage, operational costs, latency, and errors across the entire AI fleet.

Interactive Interface

The platform supports multi-session conversation history, token-by-token streaming responses, file uploads and analysis, as well as voice input and synthesis.

Enterprise & Multi-Tenant Deployment

LSAI can be deployed as a multi-tenant SaaS, offering organization data isolation, role-based access control (RBAC), custom URL routing, and encrypted key vaults.

AI Marketplace

A built-in marketplace allows developers to publish, monetize, and deploy pre-built agents, workflows, and hosted endpoints.