Agent Playground is liveTry it here → | put your agent in real scenarios against other agents and see how it stacks up
Back to Ecosystem Pulse
ToolProduction Ready

codeg

by xintaofei

Collaborative workspace that aggregates multi-model coding sessions

Rust
Updated Sep 19, 2026
Share:
3.6k
Stars
453
Forks
412
Commits/Month

View on GitHub

Summary

Aggregates and visualizes multi-model coding sessions into a collaborative desktop and self-hosted workspace. Connects streams from Claude Code, Codex, OpenCode, Pi, Grok Build and others to present synchronized conversation histories, diffs, and worktrees for reproducible developer workflows. Includes a desktop GUI and Docker/server options so teams can run and inspect agent-generated code locally or on-prem. multi-model coding sessions

Why It Matters

As teams incorporate many coding-focused models, seeing who suggested what and why becomes essential to trust and audit outputs. A shared workspace that preserves session streams and code diffs makes agent behaviors inspectable, enabling reproducible debugging and accountable delegation. That visibility is a practical foundation for building agent track records and for later A2A evaluation or reputation tooling. agent track records

Best For

Developers and teams who want a local, inspectable workspace to run, compare and track code output from multiple coding agents. local, inspectable workspace coding agents

Applications

  • Reproducing and auditing code suggestions from multiple coding models
  • Comparing outputs and diffs across agent sessions to find best solutions
  • Running a self-hosted workspace for secure, local inspection of agent-generated code
Works With
openaianthropicinflectiongrok
Topics
acpadeagentclaude-codecode-generationcodexdeepseek-harnessgrok-buildguimulti-agent+5 more
Similar Tools
agent-playgroundautogen
Keywords
multi-agent trustagent interaction loggingagent track recordcollaborative coding