duyilemi/multi-agent-workflow
A full‑stack AI application that automatically routes user requests through a supervisor‑worker agent system built with LangGraph, Groq, and Tavily. It showcases production‑grade observability, real‑time streaming, and a polished React dashboard that visualises the entire decision‑making pipeline.
SUMMARY AI summary by gpt-5-mini
Multi-Agent AI Workflow Orchestrator is a full‑stack reference implementation that routes user requests through a supervisor→worker agent system (LangGraph) to specialist agents for prompt enhancement, live research (Tavily), code execution (Python REPL) and validation. It exposes production‑grade observability, real‑time streaming (SSE) and a React dashboard that visualizes the decision pipeline. Who uses it: developers and researchers building multi‑agent orchestration, agent routing logic, or observability for AI workflows. Key features: - Supervisor routing with loop prevention and structured outputs (Pydantic). - Specialist agents: Enhancer, Researcher (Tavily), Coder (Python), Validator. - Server‑Sent Events streaming and step‑by‑step pipeline updates. - Detailed agent logs: timestamps, durations, token usage. - React+Vite frontend with live pipeline diagram, persistence, retry/backoff. - Docker Compose for local dev; deployed on Hugging Face (backend) and Vercel (frontend).
DETECTED Detected AI stack
AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.
Owner
Data Scientist | Machine Learning Engineer | Software Engineer
Dates
| Created on GitHub | 2026-05-12 |
| Last push | 2026-05-12 |
| First seen here | 2026-05-12 |
| Last fetched | 2026-05-12 16:10 |