Guides

Run AI agents in production. With your eyes open.

Practical, dated guides to observe and debug coding agents, LLM apps and automations, and to run them for customers and clients. Written by the team building Voight, from what we see in real agent runs.

By Dangel Rodriguez9 guidesLast updated Oct 11, 2026

Start here/ 01DebuggingHow to Debug AI Agents in Production (2026)Why AI agents fail without errors, and a step-by-step method to trace a bad run back to the first wrong step and fix it for good.

Dangel RodriguezUpdated Oct 11, 20266 minRead the guide →

All guides

Pick a topic. Every guide is dated and kept current.

/ 02Coding agentsClaude Code Observability: Track Every Tool Call, Token and Cost (2026)See every prompt, tool call, token and dollar of your Claude Code sessions in one timeline, with one command and your privacy level.Dangel RodriguezUpdated Oct 11, 20264 min
/ 03n8n and automationHow to Debug Failed n8n Workflows and Catch Errors in Production (2026)Find the node that broke, read its input, and set up an Error Workflow that records every failure in n8n so nothing fails silently.Dangel RodriguezUpdated Oct 11, 20265 min
/ 04DebuggingHow to Read an AI Agent Trace and Find the Root Cause (2026)A trace is the whole story of one agent run. How to read one, what a waterfall shows, and the patterns that point straight at the cause.Dangel RodriguezUpdated Oct 11, 20264 min
/ 05Coding agentsOpenAI Codex Observability with OpenTelemetry: Trace Every Session (2026)Codex exports OpenTelemetry natively. One config block turns every session into a trace with tool calls, tokens and failed API requests.Dangel RodriguezUpdated Oct 11, 20263 min
/ 06LLM appsHow to Track OpenAI and Anthropic API Costs per User (2026)Your provider bill is one number. How to attribute every LLM call to the user, plan and customer that caused it, with one line of code.Dangel RodriguezUpdated Oct 11, 20263 min
/ 07Teams and agenciesObservability for Forward-Deployed Engineers Shipping AI Agents (2026)You build the agent at the customer and then you leave. How to keep seeing what it does, prove it works, and debug it without access.Dangel RodriguezUpdated Oct 11, 20264 min
/ 08Teams and agenciesHow Automation Agencies Monitor AI Agents and n8n Workflows for Every Client (2026)Twenty clients, two hundred workflows and nobody watching at 3 a.m. A setup that catches failures per client and proves the work is holding.Dangel RodriguezUpdated Oct 11, 20263 min
/ 09MCP and deploymentDeploy and Operate AI Agents from Claude Code with MCP (2026)Deploy, message, schedule and retire autonomous agents from Claude Code, Cursor or Codex through an MCP server, with spending guarded.Dangel RodriguezUpdated Oct 11, 20263 min

See what your agents do. In one command.

Connect Claude Code, Codex or Cursor and your first traces show up in seconds.

npx -y @voightxyz/sdk setup

Start free Read the docs