DataRagAgentsNext

Munin

Ingest any data source: endpoints, databases, CSVs, PDFs. Then let workflow agents answer for you.

Role
Architect & lead engineer
Year
2026
Practice
Data
Link
munin.is
{ }APIDatabaseCSVPDFRetrieveReasonActPDF · p.4orders API

At a glance

Starting point
An agentic knowledge base + customer-service suite that ingests anything (endpoints, DBs, PDFs, CSVs) and lets non-coders ship chatbots in minutes.
What we built
Ingest any data source: endpoints, databases, CSVs, PDFs. Then let workflow agents answer for you.
Stack
Next.js, TypeScript, Pinecone, LLMs + custom reasoning, Langchain, Go, AWS

An agentic knowledge base + customer-service suite that ingests anything (endpoints, DBs, PDFs, CSVs) and lets non-coders ship chatbots in minutes.

Overview

Munin is Gott Data's AI-powered knowledge base and customer-service platform. It pairs an agentic knowledge base with a full customer-service suite (ticketing, live chat, reviews) and centralized application monitoring across multiple data sources.

The agentic knowledge base

Munin's intelligence engine understands which service to use and when. Users link endpoints for retrieval or actions, upload PDFs and CSVs, and chain custom agents into multi-step workflows. Raw data turns into actionable insights without code.

Customer service suite

  • Live chat with smart suggestions, real-time translation, and multi-language support.
  • Ticketing system with full lifecycle management.
  • Reviews and feedback aggregation.
  • Workflow automation for repetitive support tasks.

Munin application

Connects multiple digital media sources into one unified application. Centralized statistics, consolidated monitoring, actionable insights across platforms.

Stack

  • FRONTEND: Next.js
  • BACKEND / API: TypeScript
  • KNOWLEDGE BASE: Pinecone (vector DB)
  • AI ENGINE: LLMs + custom reasoning, Langchain
  • BACKEND SERVICES: Go
  • INFRASTRUCTURE: AWS
Easily one of the most sophisticated and complete projects I've shipped: ML/AI in Python, prompt engineering in Langchain, frontend in Next.js, backend in Go, infra on AWS, plus a fully functional customer service portal with a websocket system that scales.