Tencent · 2026-11-12 · notable
LoopForge — Tencent's self-hosted workflows for coding agents
In short
LoopForge is Tencent's open-source, self-hosted workspace that runs Claude Code, Codex, CodeBuddy and other agent CLIs in Docker sandboxes, through workflows that mix agents, scripts and human approval gates. MIT license.
Tencent · 2026-11-12 · 2 sources checked

Tencent's open-source workspace turns a team's delivery process into agent workflows with human sign-off.
Key specs
| GitHub stars | 1.8k |
|---|
What is it?
LoopForge gives coding agents a self-hosted home: you describe a task to Main Bot in a chat, and it works in a Docker sandbox with your repositories, hands parts to project roles and can open a pull request. Proven ways of working are saved as SOPs. The latest release, v0.3.8, renders sandbox dev servers in a side browser and asks before handing a requirement to an SOP.
How does it work?
An SOP is a fixed flow of nodes: agent roles, deterministic scripts, sub-SOPs, previews and human gates. A typical flow clarifies, designs, implements, has an independent agent review, runs tests as a script, then waits for a person to accept or send the work back. LoopForge drives existing agent CLIs such as CodeBuddy, Codex, Claude Code, OpenCode, Cursor CLI and Kimi Code rather than shipping its own model.
Why does it matter?
Teams that already pay for several coding agents get one place to run them on their own servers, with fixed review and approval steps instead of ad-hoc prompting. A LoopForge Local mode brings the same workflows straight into Claude Code, Codex, CodeBuddy or Cursor without the platform.
Who is it for?
engineering teams running several coding agents
Try it
git clone https://github.com/Tencent/LoopForge.git && cd LoopForge && ./start.sh