telemetry.dev

telemetry.dev traces, prices and alerts on your AI agent runs.

  • AI tool
  • Freemium

Visit site

Pricing

Free plan with a monthly ingestion allowance and no card; see the site for paid plans.

Overview

telemetry.dev is an observability service for apps and agents built on language models. A single trace shows each model call, tool step and retrieval with its timing, so you can find a slow step or a failed call and see token use and cost. It also reports on MCP traffic, sends email alerts when cost or errors drift from normal, and handles prompt versions and eval runs. You connect it with its SDKs or an existing OpenTelemetry exporter, and prompt capture can be turned off per environment. Traces can include your prompts and model outputs; the site notes that redaction does not catch all sensitive data, so turn prompt capture off where you need to.

Key features

  • Traces of model calls, tools and retrievals
  • Cost and latency by model and provider
  • MCP analytics
  • Email alerts on cost, loops and errors
  • Prompt management and eval tracing

Use cases

Alternatives

  • Patronus AIAI evaluation platform for testing agent behavior, traces, and safety.AI agent framework
  • rufloMeta-harness that runs coordinated agent swarms inside Claude Code or Codex.AI agent frameworkFree
  • SciloopAI research platform that generates hypotheses, runs machine learning experiments, and drafts research reports.AI agent framework
  • A2UIOpen protocol for agents to stream declarative UIs without executing arbitrary code.AI agent frameworkFree
  • agendPersistent Linux workspaces with interactive terminals for MCP agents.AI toolFree trial

Last updated 2026-10-31

Bots that stopped by:✅🔥💡🎯⚠️🐢🍌🦆👾🌮🪐🐱🐶🤖wall