Observability and cost tracking for Kimi CLI

See every Kimi CLI session, turn, tool call, subagent, and token as it happens. ClawMetry finds Kimi on your machine on its own and renders sessions, transcripts, token splits, and cost in one dashboard. Kimi Code CLI is the same runtime here, so nothing splits when you switch.

What ClawMetry sees on Kimi CLI

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

• Sessions per project: one row per Kimi session, labelled with the directory you ran it in, when it started, when it last moved, message count, and its token totals.
• The full transcript: your prompts, the model's answers, its thinking, and every tool call with its arguments and result, in the order they happened.
• Token splits that include the cache: fresh input, cached input, cache writes, and output, counted per step and added up per session, so a long agentic turn is not reported as one call.
• Subagents as their own sessions: each agent Kimi launches nests under the turn that started it, with its own tokens, tool calls, status, and the type it ran as.
• Compactions and retries: when Kimi compacts its context or retries a failed step, that is on the timeline instead of being invisible.
• Approvals and blocked hooks: what Kimi asked permission for, what you answered, and what a hook stopped.
• Cost, labelled honestly: tokens are exact, and the dollar figure is priced from your configured model and marked estimated, because Kimi does not record which model produced a session. Where a session cannot be priced honestly, ClawMetry shows the tokens and says so.
• One runtime across the rename: Kimi CLI and the standalone Kimi Code CLI land in the same view, with the same numbers.

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 finds Kimi on the machine and starts reading immediately. No config, and no need to restart Kimi.

3. Open the dashboard

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

Start the free trial →

The one thing this catches that Kimi alone does not

Kimi shows you the session you are sitting in. One turn can run for many steps, launch subagents, compact its own context, and retry a step that failed, and the terminal scrolls past all of it. ClawMetry keeps the whole fleet in one place: cost per hour across every session and project, subagents billed separately instead of hidden inside a parent, repeated-call detection, and token-growth curves per session. Set a threshold and the approval queue pauses before the next call. Kill switch (Pro) stops it.

Starred by the people building AI

Engineers and founders from OpenAI, Google, PostHog and more have starred ClawMetry on GitHub. One of them helped build it.

All 22 runtimes ClawMetry observes

Kimi CLI is one of 22. Free on every plan: OpenClaw, NVIDIA NemoClaw. Paid on Cloud Pro or Self-Hosted Pro: the rest.