Self-hosted multi-agent AI platform. Persistent labs, 40 sandboxed tools (code, web, RAG, GPU dispatcher, web3), private Qdrant + LightRAG, sandboxed code execution per lab, anti-loop detection, Ollama/vLLM/OpenAI/Anthropic adapters. Two-command Docker deploy. Apache 2.0.
react docker self-hosted orchestration multi-agent sandboxing ai-agents on-premise rag fastapi ai-platform agent-framework qdrant llm vllm ollama private-ai agentic-ai gpu-orchestration anti-loop
-
Updated
Jun 9, 2026 - Python