DEV Community: Dennis Zimmer

I built AgentMon Live to keep track of Claude Code. Now it has a browser view.

Dennis Zimmer — Fri, 06 Nov 2026 15:34:37 +0000

Running a few Claude Code sessions at once is great until I forget which one is doing what.

One is working through a change. Another has subagents running. A third might be waiting for me, but I won't know until I check that terminal again.

I kept bouncing between them to answer a few basic questions: What's happening? Does anything need me? Is the work actually finished?

That's why I built AgentMon Live.

It's a Claude Code mod that lets you watch Claude and its subagents work, see costs and usage limits, and spot when something needs your attention. It also checks for unfinished work before Claude says "done".

I shared an earlier release recently. This update adds the part I've been most excited about: you can now follow your agents live in your browser.

A browser view across your sessions

You keep working in Claude Code as usual. Open the browser view alongside it, and you can follow agent activity in real time and switch between sessions running in different terminals.

For my workflow, that last part matters a lot. Checking on several sessions means remembering where each one lives, switching terminals, and finding the relevant output. Having a browser view gives me somewhere to look while the work carries on.

The pane beside the chat is still available too. If you're working in one session, you might prefer keeping everything together. When several sessions are going, the browser gives you another way to follow them.

It's a simple addition to describe, but it's the one I've really wanted to use myself.

What I want to know while an agent works

Seeing activity is useful. Knowing when to pay attention is more useful.

AgentMon brings together running tool calls, subagent activity, changed files, and verification results. It also shows costs and how close you are to your usage limits.

That helps with the questions I keep coming back to:

What is Claude working on right now?
What are its subagents doing?
Is anything waiting for my attention?
How much of my usage allowance is left?
What has actually been checked?

I want to be able to glance over, understand the situation, and get back to what I was doing.

Checking what "done" means

The other part of AgentMon I care about is the completion check.

Consider a session where Claude runs the tests successfully, makes another code change, and then reports that everything passes. The earlier test result doesn't verify the latest edit.

AgentMon's "done gate" checks the recorded work before Claude finishes. It can catch things such as failed checks, tests that became stale after an edit, and unfinished todos. When it finds unfinished work, it sends that back to Claude once.

I still review the result myself. This helps catch some of the things I'd otherwise have to find by reading back through the session.

The README explains the checks in more detail.

Trying it

In a Claude Code session, install it with:

/plugin install agentmon-live --marketplace codenotary/agentmon-live

Follow the marketplace and scope prompts. You can open the pane with:

/agentmon

It also works with Code inside Claude Desktop. For that setup, install AgentMon as a mod in Claude Desktop itself rather than through Claude Code in your terminal.

The project is open source, with installation instructions on GitHub.

If you tried an earlier release, I'd love to hear how the browser view fits into your workflow. I'm especially interested in hearing from people juggling several sessions.

What do you find yourself checking most often while your agents are working?

Monitoring AI coding agent costs and activity on your own machine

Dennis Zimmer — Mon, 05 Oct 2026 09:58:07 +0000

I run Claude Code, Cursor, Codex and a handful of sub-agents on my own machines every day.

For a long time, I knew what they cost but had little visibility into their work. If an agent felt slow or got stuck, I had to dig through separate logs to understand why.

That became harder as I started using several agents on the same project. Each kept its own session history, and there was no single place to review the activity.

We built AgentMon to monitor agents across a company. AgentMon Start brings that visibility to a desktop app for individual developers and teams.

I’m part of the team behind AgentMon. This post explains how Start collects data and what you can inspect after installation.

How collection works

AgentMon Start reads the session logs your coding agents already write. It supports more than 16 agents, harnesses and LLMs, including Claude Code, Claude Desktop, Codex, GitHub Copilot, Cursor, Gemini CLI and Goose.

There is no SDK to add to your project or proxy to configure.

The app has three local components:

A collector that parses agent logs into snapshots of token usage, costs, tool calls, errors and security findings.
A local server that stores those snapshots in SQLite and analyzes them.
A desktop dashboard for viewing the results.

The local server listens on 127.0.0.1. Closing the dashboard leaves collection running in the background.

Start with costs

The Overview shows spending alongside session counts, token usage and cache hit rate. You can inspect the last 24 hours, seven days, 30 days or all recorded activity.

The useful question is usually more specific than “How much did I spend?”

Which session used the budget? Which model was running? Did an agent spend time retrying the same operation?

The dashboard lets you move from the total to the sessions behind it. The hourly chart helps you locate a burst of activity, while the model breakdown shows how spending was distributed.

Inspect the sessions and tool calls

Sessions groups activity into threads representing continuous work on a repository.

Each thread shows the agent, cost, token usage and activity mix, including file reads, edits, shell commands and web access. Errors and security findings appear alongside the session.

In the example from our launch post, a Claude Code thread cost $1.96 and ran 26 Bash calls. Two calls failed, and the thread was flagged for a prompt-injection finding.

Those details give you somewhere to start investigating. You can review the failed calls and decide whether the agent needs another instruction, a different tool or a stopped session.

The Tools view also aggregates failures across sessions, so you can see whether one tool accounts for most of the errors.

Review a trace

When an agent stalls or produces an unexpected result, open its session trace.

The waterfall shows recorded LLM calls, tool calls and reasoning steps with timings. The side panel provides the available input, output, metadata and raw JSON.

This helps you follow the recorded sequence of events and inspect where the session went wrong.

Set alerts and review recommendations

Alerts check every 60 seconds. Depending on your plan, they can flag cost thresholds, spending spikes, error spikes, long-running sessions and unmanaged agent activity.

Recommendations examine the last 30 days for cost and reliability problems, including low cache use, repeated retries, wasted sessions and bloated prompts. Each recommendation includes evidence and an estimated saving.

Security checks flag secrets in prompts, prompt-injection attempts and dangerous shell commands. A finding gives you something to investigate; review the session context before deciding what action to take.

Local storage and optional team sharing

Activity data stays on your machine by default.

The app contacts Codenotary for sign-in, license checks and feedback you choose to send. Connecting a device to a team dashboard is an explicit step, and the app shows what will be uploaded before syncing begins.

For teams, that shared view helps answer questions about spending across people, projects and models, as well as which agents are running.

Try it on your own sessions

AgentMon Start is available for Windows, macOS and Linux. Your first three devices are free for six months, and paid plans start at $15 per user per month.

Install the app, sign in and continue working. The collector picks up existing agent sessions, and the dashboard fills in from that activity.

A useful first check is to open your most expensive session and look at its tool calls. Was the agent making progress, waiting on something or repeating failed work?

Read the full walkthrough and get started with AgentMon Start.

If you use several coding agents, how do you currently track their costs and investigate failed sessions? I’d be interested to hear what information you’re missing.