What Claude Code really puts into each project

• Every session from start to end, with the model.
• Tokens from the local transcript: input, output, cache reads and cache writes.
• Subagents as their own sessions with their own tokens.
• Active agent runtime: time spent waiting for you does not count, idle gaps count for two minutes at most.
• A title per session: your /rename, else Claude Code's own title, else a short title from the first prompt.
• Commits made in the session, with message and branch.
• The folders the session wrote to, for assigning it to the right project.

Each session is priced at the model's list price when it started, with input, output and cache priced separately, and converted to euros. On a subscription such as Claude Max or ChatGPT, you can instead split the monthly fee across your projects by token share.

On the invoice, AI work appears as cost with a markup, as a token quantity at your price, or as active runtime at an hourly rate. Passing AI costs on is part of Pro.

• Create an API key in the app and run curl -fsSL https://app.time-momentum.com/install.sh | sh on macOS or Linux (Node.js 22.14 or later). It installs the tm command line and asks for the key.
• It adds its hooks to ~/.claude/settings.json and leaves your existing hooks in place.
• Past sessions are imported right away, as far back as Claude Code keeps its transcripts (30 days by default).

Prompts, responses and code never leave your machine. Time Momentum receives the model, tokens, runtime, working folder, repository and a session title. Where a tool keeps no title, the first prompt is sent shortened to 200 characters; Time Momentum turns it into a short title and does not store the prompt. You can switch off titles and folders entirely or exclude individual folders. The capture code is public in the npm package @time-momentum/cli.