Back to directory

ClawCodex

A production-oriented, token-efficient Python rebuild of Claude Code's reference TypeScript implementation, adding persistent multi-agent teams, a session supervisor and measured token-compression of command output.

ClawCodex

ClawCodex is described as a production-oriented Python rebuild of Claude Code, ported from its TypeScript reference implementation and extended with a Python-native runtime, aiming for token-efficient operation at scale. Its multi-agent teams feature lets a session become a leader that creates named teammates who stay alive between assignments, message each other and the leader, and pull work from a shared, dependency-aware task board, while one session-scoped supervisor admits, tracks and can interrupt every subagent, foreground or background; background workers resume with their history under the same ID, and worktree isolation runs in a real Git checkout or fails before the model runs.

Its /eco token-compression feature applies deterministic filters to command output, failure-focused test summaries, stripped git/pip/npm ceremony output, log deduplication, and recoverable head-caps for long output, reducing Bash-output tokens by a measured 80 percent on one benchmark set of real command outputs, with anything trimmed still available on disk.

The project reports 80.9 percent on the verified 89-task Terminal-Bench 2.1 suite running headless on Claude Opus 5, reproducible from an open-source Harbor adapter.

More in Coding agents

See all 793
  • The City

    github.com

    A native macOS app that shows Claude Code runs as a city, with each project as a building, each team as a floor of robots, and every animation driven by a real event from the claude CLI's stream.

  • rig

    tryrig.ai

    A small, single-binary agent runtime that points at any OpenAI-compatible model, local or hosted, reads and edits a repository's code, runs its tests and remembers what it learned, with a loop implemented in a few hundred lines.

  • Jauvex

    jauvex.reindent.com

    An Electron desktop app that puts Claude Code, Codex and Grok coding-agent sessions side by side and lets you steer them by voice, with a small model handling fast routing decisions.