How do I monitor xAI Grok CLI sessions?

Install ClawMetry (pip install clawmetry). The daemon auto-detects ~/.grok/sessions and the unified event log at ~/.grok/logs/unified.jsonl, and renders every session, tool call, model, and token in one dashboard at http://localhost:8900.

Can I see what Grok actually uploaded to xAI?

Yes. Grok stages outbound repo payloads under ~/.grok/upload_queue/ before uploading. ClawMetry surfaces the queue contents so you can audit what left your machine (files, sizes, timestamps) and flag any surprisingly large payloads before the session repeats.

Does ClawMetry track Grok cost per model?

Yes. Grok routes between Grok Fast, Grok Beta, and Grok Code models; every turn is labelled with the model id and per-model spend rolls up on the Cost tab.

Is ClawMetry safe to point at a running Grok install?

Yes. Every file (unified log and per-session JSONL) is opened read-only; the session directory scan is bounded by the CLAWMETRY_FAMILY_SESSION_LIMIT env var so a huge session history cannot slow the daemon.

How much does ClawMetry cost for Grok?

Free forever on your own machine. Cloud starts at $9/node/mo (Starter) or $19/node/mo (Pro).

Observability and cost tracking for xAI Grok

See every Grok CLI session, tool call, and token as it happens. ClawMetry reads Grok's session store under ~/.grok/sessions and its unified event log, and renders sessions, tokens, cost, and outbound repo payloads (so you see what left your machine) in one dashboard.

Supported runtime Reads Global event log at ~/.grok/logs/unified.jsonl plus per-session JSONL under ~/.grok/sessions/// Read-only Free trial

What ClawMetry sees on Grok

ClawMetry's Grok adapter opens the store Grok already writes (~/.grok/logs, ~/.grok/sessions, ~/.grok/bin/grok (overrideable via CLAWMETRY_GROK_HOME)) in read-only mode and surfaces:

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: Per-turn tokens written by Grok (input, output, plus cached tokens on cache-supported Grok models).
Cost per session, per model, per hour: cache-aware where the model supports caching.
Outbound repo-payload attribution (the ~/.grok/upload_queue/ staging area, so you can audit what actually left your machine): the Grok-specific signal ClawMetry surfaces that generic dashboards do not.
Approval and kill switch: watch for token-maxing signals (message repeat, no-progress heuristic, cost-per-hour threshold), pause or kill before the next call.

Source: the adapter is at clawmetry_pro/adapters/grok.py. Every parse rule and every SQL SELECT is in that file.

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

2. Run the daemon

clawmetry

The daemon auto-detects Grok's local store and starts reading immediately.

3. Open the dashboard

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

Start the free trial →

The one thing this catches that Grok alone does not

Left alone, Grok's failure mode is a session that stages an unexpectedly large repo payload for upload and you never notice. ClawMetry watches for it: cost per hour on the Grok 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.

All 19 runtimes ClawMetry observes

Grok is one of 19. Free on every plan: OpenClaw, NVIDIA NemoClaw. Paid on Cloud Pro or Self-Hosted Pro: the rest.

OpenClaw NVIDIA NemoClaw NanoClaw PicoClaw Claude Code Codex Cursor Aider Goose opencode Qwen Code Hermes Pi Deep Agents n8n Antigravity GitHub Copilot Grok (this page) QM