Beacon
An AI platform that converts unstructured project data — Slack threads, emails, meeting notes — into validated BRD drafts automatically.
Built by Sandeep Prajapati
Next.js 14FastAPIPythonGroq LLMPostgreSQLFirebase
← Projects→→→
Overview
Beacon is a full-stack AI platform that automates Business Requirements Document (BRD) creation. It ingests raw project data from multiple sources, filters out the noise, extracts the meaningful signals, and runs a multi-agent LLM pipeline to produce a structured, professional BRD — all reviewable and exportable from a modern web UI. Built at HackFest 2.0 with a team of four.
The Problem
Writing a BRD means manually sifting through scattered Slack threads, email chains, meeting notes, and documents — then turning that mess into a structured, stakeholder-ready document. It's slow, error-prone, and easy to miss requirements buried in the noise.
Beacon takes that scattered, unstructured input and turns it into a validated BRD draft automatically.
What I Built
- Multi-source ingestion — Slack OAuth and Gmail connectors plus file uploads pull in raw project data
- Two-phase noise filter — heuristics followed by a Groq LLM batch classifier separate signal from noise
- Multi-agent BRD pipeline — specialized agents generate the structured document, with a validator and exporter
- Human-in-the-loop versioning — drafts are reviewable, editable, and collaboratively refined before export
- Modern web app — auth, sessions, ingestion UI, BRD review, collaboration, and export
Architecture
frontend/— Next.js 14 app: auth, sessions, ingestion UI, BRD review, collaboration, exportbackend/api/— FastAPI server: REST endpoints, orchestration, streaming- Noise filter module — two-phase classifier (heuristics + Groq LLM batch)
- BRD module — multi-agent generation pipeline, validator, exporter, HITL versioning
- Integration module — Slack OAuth + Gmail connectors
Tech Stack
Frontend: Next.js 14, TypeScript, Tailwind CSS
Backend: FastAPI, Python 3.9+
AI: Groq LLM engine, multi-agent classification + generation
Data & Auth: Firebase (Auth & Firestore), PostgreSQL
Use Case
Product teams, consultants, and engineering orgs who want to compress hours of requirements-gathering into minutes — turning the chatter that already exists across their tools into a clean, validated BRD.
Related projects
Noloop
A glass-box medical claims adjudication and fraud-defense platform — an AI agent pipeline turns unstructured claim documents into explainable, role-specific decisions for patients, hospitals, and insurers.
Echo Echo
An AI music generation platform — multi-agent LLM orchestration (CrewAI + Groq Llama 3.1) composes 30-second instrumental sketches via the Suno v4 API, generates thematically aligned lyrics, and scores every output through an automated copyright-similarity detection pipeline.
FinSight AI
An AI-powered financial intelligence platform — parallel agents deliver probabilistic revenue forecasts, buy/sell signals, and peer benchmarks in under 20 seconds.