agentacct

agentacct is a local dashboard for developers who want to review what their coding agents did across projects and compare usage with recorded results. It brings task activity, file changes and test outcomes together with token counts and estimated costs. Claude Code and Codex have the fullest task records; it also supports OpenCode, Kimi Code and Hermes, with coverage varying by capability.

Each task has a Work Receipt that connects participating sessions to an activity timeline, recorded outcomes and check results. An agent saying it's done counts as Reported. Verified requires passing checks after the last recorded change. Earlier failed attempts remain visible, so you can see the history behind a result.

Usage views let you filter by date, agent, model or provider and compare fresh tokens with cache reads and writes. Costs come from pricing tables, not invoices. Confidence labels mark how reliably usage connects to recorded work, and missing evidence appears as a gap. Project views gather work from different agents into a shared timeline while labeling partial coverage and overlap.

The MIT-licensed open source tool runs on macOS and Linux, with Windows support through WSL. It has a terminal dashboard and a native app for macOS 14 or later. All stored data stays on your computer, with no account, cloud service or telemetry. It records work metadata, including touched files and credential-scrubbed commands, but doesn't store prompts, model responses or conversation transcripts. It never requests or stores provider API keys.