abtop vs agentacct
htop for AI coding agents: every Claude Code, Codex and OpenCode session in one TUI — tokens, context-window %, rate limits, child processes, orphan ports. Read-only, no API keys. Rust. — versus — Local-first work receipts for coding agents: reads the session logs Claude Code, Codex, OpenCode and Hermes already write and reports what each task did, what it cost, and whether a check proved it.
Opposite time horizons. abtop is htop for agents — live sessions, context-window %, rate limits, orphan ports, right now. Agentacct is the record afterwards: what the task did, what it cost, whether it was verified.
| abtop | agentacct | |
|---|---|---|
| Stars | 3.5k | 667 |
| Forks | 315 | 75 |
| Language | Rust | Python |
| License | MIT | MIT |
| Last activity | 1 months ago | today |
| Topics | coding | coding, local |
| Curated connections | 5 | 7 |
abtop — the curator's take
The moment you run 3+ agents in parallel this earns its place: live per-session context-window bars, real-time rate-limit quota, and the sleeper feature — orphan port detection for the dev server an agent forgot to kill. Read-only from local process/file state, cross-platform, Enter jumps to the agent's terminal (tmux/cmux aware). NOT an analytics tool: it's a live dashboard, not history — for cost breakdowns, replays and budgets over time you want a session-log analyzer, not a top. OpenCode monitoring needs the sqlite3 CLI installed.
agentacct — the curator's take
The only tool in this neighbourhood that refuses to take the agent's word for it. Everything else here counts tokens and dollars; agentacct counts work and evidence, and keeps them on separate axes — an agent saying 'done' files the task under Reported, and only a machine check that postdates the newest recorded work earns Verified. Every figure wears its basis (`≈` for estimated, bare ` only for reported), every field carries provenance (client hook, transcript scan, MCP record), and unconnected verifier sources stay honestly labelled 'not connected' instead of quietly scoring zero. macOS app, `agentacct tui`, and a loopback-only JSON API; no account, no telemetry, never touches your provider key. Use it when you are accountable for agent output. Overkill if you only want to know this month's spend.