Telemetry
For teams shipping agents, tool calls, and LLM workflows

AI Agent Observability

Track agent runs, tool calls, retries, latency, model cost, failures, and accepted outcomes in SQL-ready event tables.

Get an API keyCopy the prompt
Why this works
  • Understand which agent workflows are reliable enough for production users.
  • Connect model cost, latency, and failure rate to real product outcomes.
  • Debug tool-call failures without storing raw prompts or completions.

Agent prompt

Paste this into your coding agent

Replace YOUR_API_KEY after signup, then ask the agent to run the product flow and verify the first events.

agent prompt

AI Agent Observability setup prompt

text
Instrument AI agent observability with Telemetry.

Use /skill.md and this Telemetry API key: YOUR_API_KEY

Please log agent runs, model requests, tool calls, retries, failures, and accepted outcomes. Include agent_name, feature, model, provider, status, latency_ms, retry_count, tool_name, total_tokens, estimated_cost_usd, and error_type where relevant.

Create dashboards for run volume, failure rate by tool, p95 latency, model cost by feature, and accepted-output rate.

Do not log raw prompts, completions, auth headers, secrets, or private user content.

Setup steps

  1. 1Add structured events around agent runs, model calls, and tool calls.
  2. 2Record status, latency, retry count, model, provider, and estimated cost.
  3. 3Capture user-visible outcomes like copied, saved, accepted, or retried.
  4. 4Create dashboards for failure rate, p95 latency, tool errors, and cost by feature.

Events to capture

agent_run_startedagent_tool_calledagent_run_completedagent_run_failedai_output_acceptedllm_request_completed

Questions unlocked

  • Which tools fail most often?
  • Which models produce the best accepted-output rate per dollar?
  • Where do retries or timeouts damage conversion?

Next step

Create the API key your agent will use

The free plan is enough to run the prompt, send test events, and review the first dashboard.

Prompt packStart free

Related pages

Claude Code Observability

Give Claude Code a prompt that makes telemetry part of the implementation pass instead of a separate cleanup project.

Open page

Cursor Telemetry Setup

A copyable setup prompt for adding event tables, funnel tracking, and dashboards from inside Cursor.

Open page

Codex Instrumentation Prompt

A focused prompt that asks Codex to instrument the rest of the product, verify events, and summarize coverage gaps.

Open page