Agent infrastructure
agentacct
mikehasa/agentacct
A local-first work ledger for coding agents that breaks tasks into steps, tools, files, tests, tokens, and estimated cost, separating reported outcomes from verified ones.
Topics
Related projects
hive583A local-first multi-agent workbench that runs Claude Code, Codex, Gemini, OpenCode, and other CLI agents as real PTY processes, with an orchestrator delegating tasks through a shared task graph.
herdr49.7kA terminal multiplexer built for coding agents with persistent sessions, multi-machine views, and agent status tracking.
herdr48.6kA terminal runtime and workspace manager for coding agents, keeping sessions alive across detaches, organizing multiple machines, and exposing panes to agents through a CLI and socket API.
codeburn11.7kA local observability tool that tracks AI coding token spend across dozens of agents, broken down by model, project, and task.
abtop3.8kabtop is a btop-style terminal monitor for AI coding agents, showing Claude Code, Codex and OpenCode sessions with token usage, context window, rate limits, child processes and open ports.
agent-deck1.3kAgent Deck is a terminal control center for AI coding agents. One TUI manages Claude, Codex, OpenCode, and other CLI sessions with status detection, groups, forking, worktrees, MCP management, and cost tracking.