Skip to content

awesome-ai.md

Latest commit

 

History

History
61 lines (43 loc) · 3.72 KB

awesome-ai.md

File metadata and controls

61 lines (43 loc) · 3.72 KB
Raw

Add JuggleNet6 - AI agent workflow platform with image/video generation

JuggleNet6

Release .NET 8 Vue 3 License

An AI-powered workflow orchestration platform — combine natural-language chat with executable flow diagrams. Describe a requirement in chat and watch the LLM generate a complete API orchestration with tool calling, image/video generation, and multi-turn reasoning. Self-hostable, multi-tenant, Docker-ready.

GitHub: https://github.com/pythonHuang/JuggleNet6

Tags: ai-agent | llm-app | workflow | multimodal | function-calling | rags | net8 | vue3

Highlights

  • Conversational Flow Orchestration: describe requirements in natural language; the LLM generates complete workflow diagrams with METHOD nodes, condition branches, loop structures, and variable mappings — no manual drag-and-drop needed
  • Function Calling / Tool Use: attach existing APIs and flows as function-calling tools; the LLM calls them on demand (up to 4 rounds), feeding results back for the final answer
  • Knowledge Base (RAG): upload documents (Word / Excel / PDF / TXT / Markdown), auto-chunked and embedded; AI nodes perform keyword + vector cosine-similarity retrieval for grounded Q&A
  • Image Generation: text-to-image via DashScope (通义万相) async tasks or OpenAI-compatible images/generations endpoint; results rendered inline with download/open buttons
  • Video Generation: text-to-video via DashScope async tasks with polling (up to 3 min timeout); rendered in chat bubbles
  • Custom AI Assistants: configure system prompts, input/output params, quick-prompt buttons; multi-turn chat with full history and structured JSON outputs
  • AI-Powered Report Generation: chat to describe report needs; the LLM designs datasets, query parameters, and A4 layouts with formulas (SUM, IF, etc.) and live HTML preview
  • Multi-modal Input: AI nodes accept image inputs (data URLs) for vision model description and file-parse nodes
  • Self-hosted & Multi-tenant: JWT-based auth, tenant-scoped data isolation, Docker one-click deploy

Supported LLM Providers

OpenAI-compatible interface — works with DeepSeek, Qwen (通义千问), Kimi, OpenAI, GLM, Ollama, and any custom endpoint.

Quick Start

docker run -d \
  --name juggle \
  -p 9127:9127 \
  -v juggle_data:/data \
  pythonhuang/juggle-net8:v1.0

Access at http://localhost:9127 — default account: juggle / juggle

Tech Stack

Layer Technology
Backend ASP.NET Core 8 / EF Core 8 / SQLite
Frontend Vue 3 / Vite / Element Plus / Pinia
AI Runtime OpenAI-compatible SDK / DashScope
Embeddings Vector cosine-similarity search
Container Docker (multi-stage build)
Auth JWT + RBAC Role-Based Access Control

Why I recommend it

JuggleNet6 is one of the most complete open-source AI agent platforms I've seen: it goes beyond simple chatbots by letting you turn conversational requirements into executable workflows with real tool calling, RAG knowledge bases, and multimodal generation (image/video) — all self-hostable with a single Docker command.