OpenOPC: Open-Source Framework to Build Autonomous AI Companies
aigclink · x · 2026-08-11
The HKUDS team from HKU open-sourced OpenOPC, a framework designed to let AI autonomously build and run an "AI-native company."
Core Mechanics:
- Self-Built: Automatically creates organizational structures, recruits AI employees for various roles, and establishes reporting lines based on a given goal.
- Self-Run: Drives task distribution via a state machine. Managers decompose, assign, and review tasks with support for parallel execution and dependency queuing, only escalating blockers to humans.
- Self-Grown: Attributes execution traces and user feedback to specific roles, distilling them into private experiences. Repeatedly validated experiences are upgraded into shared playbooks for new agents to inherit.
Execution & Distribution:
The execution layer integrates with mainstream coding tools like Codex, Claude Code, and Cursor. The entire AI company can be packaged into a .opcpkg format for distribution. The framework claims coverage across nine core scenarios, including software development, finance, sales, and e-commerce.
Related event: HKU Open-Sources OpenOPC for Fully Automated AI Companies(2 posts)→
More from coding & agent
- CMU Launches ExploitBench: Testing AI Agents on Real V8 Exploitation — cyb3rops · 2026-08-11
- Zhipu's ZCode Hits 1M Users: GLM Coding Plan Resets Limits, Boosts Multi-Agent Collaboration — jietang · 2026-08-11
- 4x DGX Spark Cluster Achieves 44.6 tok/s on GLM-5.2 for Real Agent Workloads — EAccelerate_42 · 2026-08-11
- Developer Calls for AI Agent Platforms to Expose Cost Control APIs — arthurcolle · 2026-08-11
- Why Are AI Agents Still Fragile in Real-World Workflows? — No_Progress92 · 2026-08-11
- Agent Memory Distillation boosts small LLM agents by up to 27.2% accuracy via hierarchical teacher memory — kaist-ai · 2026-08-11