How do I monitor Codex sessions?

Install ClawMetry (pip install clawmetry, or the one-liner installer at clawmetry.com/install.sh). The daemon auto-detects Codex and renders every session, tool call, model, and token in one dashboard at http://localhost:8900. No SDK, no code change.

Does ClawMetry track Codex reasoning-token cost?

Yes. Codex's o-family models bill reasoning tokens on top of output tokens; ClawMetry splits reasoning from output and prices them separately on the Cost tab. You can see per-session reasoning-token cost and catch a run that's silently reasoning its way through your budget.

How do I see which model Codex chose for a turn?

Every turn in the Brain tab is labelled with the exact model id (gpt-4o, gpt-4o-mini, o3-mini, etc). Model routing is visible per turn and rolled up per session, so you can see when a session flipped to a more expensive model and whether it was worth it.

Can ClawMetry catch a Codex run that goes into a runaway loop?

Yes. ClawMetry watches per-turn token growth and message-repeat signals. When a session is token-maxing (a runaway loop) the approval queue flags it and the kill switch (Pro) stops the session before the next call. Same guard covers Codex autonomous mode.

How much does ClawMetry cost for Codex?

Free forever on your own machine (local dashboard at http://localhost:8900). ClawMetry Cloud starts at $9 per node per month (Starter — observability across all 18 paid runtimes including Codex) or $19 per node per month (Pro — adds policy gates, evals, per-run waste flags).

Codex × ClawMetry

Observability and cost tracking for OpenAI Codex

See every Codex session, tool call, and token as it happens. ClawMetry detects Codex automatically and renders sessions, tokens, cost, and model routing in one dashboard. No SDK.

Supported runtime Zero config, auto-detected Read-only Free trial Official Codex site ↗

📦 631k+ installs 🌍 128+ countries ⭐ 392 on GitHub 🏆 #5 Product Hunt

What ClawMetry sees on Codex

Once ClawMetry is running, your Codex activity shows up as:

Sessions: started at, last activity, model, message count, per-session token totals.
Live activity: every tool call as it lands, with args and result. Turn-level tokens split.
Per-turn tokens: OpenAI-shaped usage block (prompt_tokens, completion_tokens, plus cached tokens when routed to a cached model).
Cost per session, per model, per hour: cache-aware where the model supports caching. Compare Codex spend against the other 19 runtimes ClawMetry observes.
Model routing across GPT-4o, GPT-4o-mini, and o-family reasoning models: the Codex-specific signal ClawMetry surfaces that generic dashboards do not.
Approval + kill switch: watch for token-maxing signals (message repeat, no-progress heuristic, cost-per-hour threshold), pause or kill the session before the next call.

Enable it in about two minutes

1. Install ClawMetry

Pick the line that matches your OS. Windows users: paste the install.cmd line, not the shell one.

# macOS or Linux
curl -fsSL https://clawmetry.com/install.sh | bash

# Windows (cmd.exe or PowerShell)
curl -fsSL https://clawmetry.com/install.cmd -o install.cmd && install.cmd

Already have Python? pip works too; onboard is the interactive first-run wizard.

# Any OS with Python 3.9+
pip install clawmetry && clawmetry onboard

Or skip the terminal entirely and grab the desktop app: macOS (.dmg) · Windows (.exe) · Linux (.AppImage). It installs and updates ClawMetry for you.

2. Run the daemon

clawmetry

The daemon auto-detects Codex and starts reading immediately. No env var, no config, no runtime restart.

3. Open the dashboard

Open http://localhost:8900. Switch the runtime chip to Codex.

Start the free trial →

The one thing this catches that Codex alone does not

Left alone, Codex's failure mode is long autonomous runs that quietly rack up reasoning-token cost. ClawMetry watches for it: cost per hour on the Codex runtime, message-repeat detection, and per-session token-growth curves. Set a threshold, and the approval queue pauses before the next call. Kill switch (Pro) stops it.