whereismyai
Agentic Telemetry
Agentic Telemetry is a lightweight, local-first viewer for OpenTelemetry traces that visualizes agentic system executions.
Agentic Telemetry is a lightweight, local-first viewer for OpenTelemetry traces from agentic systems. It ingests OTLP/HTTP JSON trace requests through a dedicated endpoint and stores trace details locally in an SQLite database file. A web dashboard renders parent-child span trees alongside timing waterfalls, automatically highlighting agent calls, handoffs, LLM calls, tool calls, and error spans. The system automatically refreshes trace listings every 10 seconds and supports span updates from exporter retries. It is built for developers running local workflows without requiring distributed setups or external trace services.
Agentic Telemetry at a glance
Key features
- Ingest OTLP/HTTP JSON trace requests via POST /v1/traces
- Render parent-child span trees with timing waterfalls
- Highlight agent calls, handoffs, LLM calls, tool calls, and errors
- Infer agent and tool labels from standard genai attributes
- Refresh the trace dashboard automatically every 10 seconds
- Store trace data locally in a telemetry.db file
- Query trace summaries and specific spans via GET endpoints
- Seed demo traces demonstrating handoffs and tool executions
Use it for
- Inspect and debug multi-agent handoffs during local development
- Analyze tool call sequences and execution timing in agentic runs
- Review LLM call performance and trace errors in parent-child trees
- Receive OTLP/HTTP JSON exports directly from local OpenTelemetry collectors
Alternatives to Agentic Telemetry
- Agentic Bug Hunter — It drives the reconnaissance, triage and reporting loop an agent would run.
- New Relic — Unified visibility across applications, infrastructure, and AI models.
- VoltAgent — Offers open-source flexibility with built-in supervisor agent orchestration and RAG.
- flightdeck — Monitor and control production and coding agents with real-time token budgets and kill switches.
- AG-UI — Standardizes agent-user interactions and generative UI for applications.
- TraceRoot — Closes the self-improving loop for coding agents using production traces and evals.
- OpenCode — Works with 75+ model providers, local models, and existing subscriptions
- OpenLLMetry — Integrates into applications with just two lines of code.
- AgentsView — It keeps your AI coding session archive local by default while tracking costs and activity.
- Polylane — It autonomously investigates production errors and submits ready-to-merge PR fixes.
- Raindrop Workshop — Provides real-time observability and self-healing evaluation loops for local AI agents.
- AgentScore — Monitors production agents and dispatches coding agents to fix failures automatically.
All Agentic Telemetry alternatives