wtf is wlog?

Wlog is a small local dashboard that shows what your Claude Code AI coding sessions have actually been doing: how many tokens they used, how much they cost, and which tools they called and whether those calls were approved or rejected. It reads the session history that Claude Code already writes to a folder on your own computer, so there is nothing to send anywhere else. The tool ships as a single program with no installer steps beyond copying it to your machine. Running it starts a small local web server with a built in interface, and your past sessions appear immediately with no setup required. There is also an optional live mode: turning on OpenTelemetry, a standard way of streaming activity data, adds real time updates while a session is running and shows whether each tool call was accepted or rejected and by what source. The dashboard has six screens. A Now screen shows the current session's cost and active tool call as it happens. A Daily screen shows a calendar style grid, shaded by how much was spent or how many tokens were used each day. A Cost and Tokens screen shows spending over time and a breakdown by model. A Sessions screen lists every past session with its cost, duration, and token totals. A Tools screen summarizes how often tool calls were approved or rejected. A Timeline screen shows the full sequence of events within one session, from the prompt through each tool call and its result. The README contrasts this with heavier options like sending the same data to a cloud monitoring service or running a stack of several separate programs together just to view usage charts. Wlog is meant to be simpler: install one file and see a working dashboard within a few minutes. It can be installed with the Go programming language's install command, or by downloading a prebuilt file from the project's releases page. A separate desktop app version is also available, built with the Tauri framework, which bundles the same program inside a normal windowed application.