Welcome to the VOID AI organization.
VOID is an open-source platform for observing, understanding, and improving AI agent systems. It builds on OpenTelemetry and observability tools like SigNoz to transform raw telemetry into actionable engineering insights.
| Repository | Description |
|---|---|
| void | Project overview, architecture, roadmap, and documentation. |
| void-sdk | OpenTelemetry instrumentation SDK for AI agents. |
| void-server | Incident intelligence engine, evaluation, fingerprinting, and reporting APIs. |
| void-dashboard | Web dashboard for incidents, replay, reports, and system insights. |
| void-demo | Sample AI agents with intentionally injected failures for demonstrations. |
| void-examples | Integration examples for LangGraph, CrewAI, OpenAI Agents SDK, Google ADK, and more. |
| void-docs | Architecture decisions, research, diagrams, and design documents. |
Modern observability platforms can tell you what happened.
VOID focuses on answering:
- Why did it happen?
- Is it important?
- Has this happened before?
- Which traces belong to the same incident?
- What should engineers fix first?
Rather than replacing existing observability tools, VOID builds on top of them to provide incident intelligence for AI agents.
- OpenTelemetry
- SigNoz
- TypeScript
- Node.js
- React
- PostgreSQL
- Redis
🚧 Early development
This project is being actively developed and the architecture will continue to evolve.
Built with ❤️ for AI Agent Observability.