[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:agentacct":3},"\u003Ch1>agentacct\u003C\u002Fh1>\n\u003Cp>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmikehasa\u002Fagentacct\u002Factions\u002Fworkflows\u002Ftests.yml\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fgithub.com\u002Fmikehasa\u002Fagentacct\u002Factions\u002Fworkflows\u002Ftests.yml\u002Fbadge.svg\" alt=\"tests\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fagentacct\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fagentacct.svg\" alt=\"PyPI\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fagentacct\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fpyversions\u002Fagentacct.svg\" alt=\"Python\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmikehasa\u002Fagentacct\u002Fblob\u002FHEAD\u002FLICENSE\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-yellow.svg\" alt=\"License: MIT\" \u002F>\u003C\u002Fa>\u003C\u002Fp>\n\u003Cp>\u003Cstrong>See what your coding agents actually did — and whether you can trust it — across Claude Code, Codex, OpenCode, and Hermes, without any of it leaving your machine.\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cp>agentacct is local-first Agent Work Intelligence for coding agents. It reads the session logs your agents already write on disk — Claude Code, Codex, OpenCode, and Hermes — joins them with the work each session records as it goes, and turns the result into one honest \u003Cstrong>Work Receipt\u003C\u002Fstrong> per task: what it did (the commands it ran, the files it touched, the tools it used), what it cost, and how well that is actually proven. Each receipt reads like an audit record, not a vibe: the decision (\"the agent says it's done\") and the evidence (\"a machine check proves it\") are separate axes, and every evidence tier has its own shape — an agent's claim can never dress up as verification. See it in the \u003Cstrong>macOS app\u003C\u002Fstrong>, a live terminal dashboard (\u003Cstrong>\u003Ccode>agentacct tui\u003C\u002Fcode>\u003C\u002Fstrong>), or over a local JSON API. No browser tab, no hosted server, no account.\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fmikehasa\u002Fagentacct\u002Fmain\u002Fdocs\u002Fassets\u002Fapp-work-receipt.png\" alt=\"A Work Receipt in the macOS app — a Verified task with its summary strip (actions, estimated cost, elapsed, 4\u002F4 checks passed), the receipt dimensions ledger with per-field provenance chips, the live checks with exit codes, and the evidence-coverage card\" \u002F>\u003C\u002Fp>\n\u003Cp>\u003Cstrong>Private by design.\u003C\u002Fstrong> Everything stays on your machine: state is plain local files, the only listener is a loopback-only local JSON API (\u003Ccode>127.0.0.1\u003C\u002Fcode>) that onboarding starts and \u003Ccode>agentacct stop\u003C\u002Fcode> stops, and there is no phone-home telemetry, no account, no cloud sync. agentacct never stores or requests a provider API key.\u003C\u002Fp>\n\u003Cp>\u003Csub>Screenshots show a synthetic demo workspace; your own dashboard renders your machine's real local data.\u003C\u002Fsub>\u003C\u002Fp>\n\u003Ch2>What you get\u003C\u002Fh2>\n\u003Cp>The same Task-primary view of your agents' work in the macOS app, in \u003Cstrong>\u003Ccode>agentacct tui\u003C\u002Fcode>\u003C\u002Fstrong> (a live terminal dashboard), and over a local JSON API — everything at a glance across all four agents, in light and dark, with a menu-bar glance always one click away:\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fmikehasa\u002Fagentacct\u002Fmain\u002Fdocs\u002Fassets\u002Fapp-dashboard.png\" alt=\"agentacct — the macOS app dashboard: recent work with decision badges and evidence-tier pips, a needs-review card, live active work, per-agent plan meters, and the daily fresh-token history for the whole workspace\" \u002F>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cp>\u003Cstrong>A receipts workbench.\u003C\u002Fstrong> Every task your agents touch becomes a row you can hold them to: lifecycle tabs that never inflate a claim (an agent's \"done\" files under \u003Cem>Reported\u003C\u002Fem> — \u003Cem>Verified\u003C\u002Fem> is reserved for machine-checked completion), an evidence column whose pip shape carries the tier, a checks column with real pass\u002Ffail tallies, and a cost column where every figure wears its basis (\u003Ccode>≈\u003C\u002Fcode> marks an estimate — a bare \u003Ccode>$\u003C\u002Fcode> is reserved for reported figures). Sorted latest-first, with an attention-first sort one click away when the one blocked task should outrank nine finished ones.\u003C\u002Fp>\n\u003Cp>\u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fmikehasa\u002Fagentacct\u002Fmain\u002Fdocs\u002Fassets\u002Fapp-work-table.png\" alt=\"The Work receipts table — lifecycle tabs (Attention \u002F Verified \u002F Reported \u002F In progress \u002F Observed \u002F Stopped), evidence-tier pips with checked ratios, per-client chips, a checks-passed rail, estimated costs, and recency\" \u002F>\u003C\u002Fp>\n\u003C\u002Fli>\n\u003Cli>\u003Cp>\u003Cstrong>One Work Receipt per task — what it did, and whether you can trust it.\u003C\u002Fstrong> Open a row and the task reads like an audit record (the screenshot at the top): what it was, who ran it, the \u003Cstrong>actions\u003C\u002Fstrong> it took (the commands it ran and the files it touched — read straight from each agent\u003C\u002Fp>\n\u003C\u002Fli>\n\u003C\u002Ful>\n",1787958252789]