lb08111/nexora-ai-platform

Nexora AI Platform is an enterprise workspace where multiple team members can share AI agents under controlled access and with full audit trails. It is built on top of QwenPaw, an open-source multi-agent framework, and adds the access control, governance, and usage tracking layers that organizations need before deploying AI tools internally. The base QwenPaw layer includes AI agents that hold conversations, run scheduled tasks, connect to messaging platforms such as DingTalk, Feishu, WeChat, Discord, Telegram, Slack, and others, and call external tools or custom skills. Agents can use cloud AI providers including OpenAI, DeepSeek, and Claude, or run models locally through tools like Ollama or LM Studio. A built-in code editor with a file tree, tabbed interface, and diff review is also included. Nexora adds enterprise features on top: a two-role system with admin and operator accounts, per-user agent grants that control which AI agents each person can access, an approval workflow for installing new tools or skills that flags high-risk actions for admin review, and a full audit log stored in PostgreSQL that records authentication events, chat messages, tool calls, and configuration changes. Token usage per user, agent, and model is tracked and shown in a dashboard. The architecture keeps Nexora-specific code in separate directories from the upstream QwenPaw core so that core updates can be merged without conflict. The frontend is built with React and Vite. The backend runs on FastAPI with PostgreSQL 16 for all enterprise data storage. The README describes the platform as a central hub connecting intelligent agents with the people and tools that run a business, under controlled governance. The full README is longer than what was shown.