An agentic AI platform built for SEBI Compliance & SME IPO DRHP Generation — enabling SME promoters to generate a substantially complete, SEBI ICDR-compliant Draft Red Herring Prospectus (DRHP) — reducing preparation time from months to hours, with zero specialist legal knowledge required.
- 💡 The Problem & The Solution
- ✨ Core Features Breakdown
- 🖥️ Interactive Dashboard & User Experience Flow
- 🛠️ Technology Stack
- 🏗️ System Architecture & Visual Pipeline
- 📂 Detailed Project Structure
- 🚀 Getting Started
- 🤝 Contributing to IPO Sahayak
- 📄 License & Attribution
Preparing a Draft Red Herring Prospectus (DRHP) under SEBI ICDR Regulations is a manual, expensive bottleneck for India's 500,000+ IPO-eligible Small & Medium Enterprises (SMEs):
- 💸 Cost Barrier: Preparing a DRHP costs ₹15–25 Lakhs in advisory and legal fees.
- ⏳ Time Bottleneck: Takes 4–6 months of continuous manual audit, drafting, and cross-verification.
- 📈 Capacity Limit: High fees and advisory complexity restrict listings to only 150–200 companies annually across India.
IPO Sahayak solves this bottleneck through an automated, plain-language wizard and intelligent AI document engine:
- 🤖 Auto-Population & Extraction: Integrates directly with MCA21, GSTN, and Account Aggregator APIs while using LlamaParse & PyMuPDF to auto-fill 70% of necessary company data.
- 📑 Exhaustive Generation: Instantly compiles 200+ page Draft Red Herring Prospectus (DRHP) documents alongside a 10-page Abridged Prospectus using DeepSeek-V4-Flash and DeepSeek-V4-Pro models.
- ⚡ Drastic Impact: Cuts document preparation costs by 60–70% (down to ₹5–8 Lakhs) and slashes timelines by 70% (down to 4–6 weeks).
- DeepSeek-V4-Flash: Rapidly translates raw business descriptions, promoter profiles, and financial notes into formal legal clauses aligned with SEBI ICDR Schedule VI guidelines.
- DeepSeek-V4-Pro: Dynamically designs visual PDF layouts, statutory header hierarchies, and complex tabular structures for financial statements.
- 5-Section AI Classifier: Automatically tags and categorizes input disclosures into Corporate Identity, Business Overview, Financial Dossiers, Risk Factors, and Legal Litigations.
- BM25 Lexical Keyword Retrieval: Matches exact statutory terms, act references, and regulatory sections from SEBI ICDR guidelines.
- FAISS kNN Semantic Retrieval: Uses
all-MiniLM-L6-v2dense vector embeddings to perform semantic similarity searches over past DRHP disclosures. - Reciprocal Rank Fusion (RRF): Fuses lexical and semantic search ranks to extract the top-k most relevant legal contexts into the AI prompt window.
- Multi-Format Ingestion: Extracts text and structured tables from scanned PDFs, balance sheets, CA certificates, and Memorandums of Association (MoA).
- Financial Table Parsing: Extracts key financial metrics—including Revenue from Operations, EBITDA, Restated PAT, and Net Worth—and formats them into normalized JSON schemas.
- Rule Check Automation: Runs real-time validation checks against SEBI eligibility norms (e.g., minimum net worth, 3-year profitability criteria, promoter dilution limits).
- Discrepancy Alerts: Highlights missing statutory disclosures, unverified promoter PAN entries, or unhedged risk factors before document finalization.
- Dual-Panel Split Viewer: Side-by-side view comparing source documents (scanned ITRs/audits) with generated DRHP pages.
- In-Browser Annotation (
react-pdf+PDF.js): Allows Merchant Bankers, Chartered Accountants, and Legal Counsel to highlight text, leave threaded review comments, and affix digital signatures.
- Celery + Redis Queues: Offloads PDF compilation, WeasyPrint/ReportLab layout rendering, and batch AI queries to background workers.
- Brevo Email Service: Triggers transactional emails, security verification codes, team invitation links, and review handoff notifications.
Experience the intuitive, step-by-step SME promoter & Merchant Banker user flow of IPO Sahayak—from initial CIN onboarding registration to AI document auto-extraction, SEBI compliance validation, collaborative review with digital text highlighters, and Brevo email dispatch.
The portal entry screen welcoming SME Issuers and Merchant Bankers. Promoters choose between registering a new SME Issuer account or logging in to an existing filing session.
Promoters enter their 21-character Corporate Identification Number (CIN). The platform queries MCA21 databases, instantly verifies company status (e.g. Reliance Jio Infocomm Limited), and auto-fills corporate information, primary email, and security verification credentials.
The main command dashboard displaying real-time application completion progress (67%), document upload counts (27/35 docs), SEBI compliance score (82%), and advisor review phase. Includes DeepSeek-V4 live multi-language translation (English ↔ Hindi).
An adaptive, plain-language form wizard that collects statutory documents and auto-extracts business models, financial highlights, and risk factors using LlamaParse & PyMuPDF.
📸 Click to view 8-Panel DRHP Wizard Screenshots
Upload Certificate of Incorporation (COI). The system automatically parses company name, CIN, incorporation date, and registered address.
Ingest 3-year restated financial statements and auditor reports. AI automatically extracts key financial KPIs including Revenue from Operations, PAT, EBITDA Margin, and Net Worth.
AI pre-toggles active revenue channels (B2B/B2G direct sales, distributors, e-commerce storefronts) based on uploaded contracts and calculates client dependency metrics.
An automated regulatory compliance engine checking statutory requirements under SEBI ICDR Regulations (2018) Chapter XB. Displays live audit ratings (Excellent Standing - 67 Score), category progress, and priority alerts for promoter contribution rules.
A secure document repository managing uploaded PDFs, SHA-256 digital signature validation, and audit history logs. Includes an in-browser digital text highlighter allowing legal reviewers to annotate source financial statements with yellow/red audit highlights.
Grant role-based filing access to Chartered Accountants, CFOs, and Legal Counsel. Dispatches shared DRHP documents directly to team members via Brevo's email notification service.
Instant compilation of the official 200+ page Draft Red Herring Prospectus (DRHP) and 10-page Abridged Prospectus with Jio branding overlay, PDF preview controls, and one-click download options.
| Layer | Technologies & Frameworks |
|---|---|
| Frontend | Next.js 15, React 18, TypeScript, Tailwind CSS, react-pdf, PDF.js, Zustand / Context API |
| Backend API | Python 3.10+, FastAPI, Pydantic, REST API, Uvicorn |
| Database & Caching | PostgreSQL (Relational Data), SQLite3 (Metadata), Redis (Cache & Celery Broker) |
| Task Queue & Async | Celery (Background Worker Queue) |
| Search & Retrieval Engine | Elasticsearch 8+, FAISS Vector Index (BM25 + kNN Hybrid Search, RRF Fusion) |
| AI & Machine Learning | DeepSeek-V4-Flash, DeepSeek-V4-Pro, LlamaParse, all-MiniLM-L6-v2 Embeddings |
| PDF Extraction & Parsing | PyMuPDF (fitz), pdfplumber, LlamaParse |
| PDF Compilation & Layout | WeasyPrint, ReportLab, PyPDF2, pikepdf, LaTeX Compiler |
| Communications & Dispatch | Brevo API (Transactional Emails, Collaboration Invitations, Notifications) |
| Government Integrations | MCA21 API, GSTN API, Income Tax PAN Verification API |
The entry flow for Merchant Bankers and SME Promoters. Users upload initial corporate documents (MoA, CA certificates, ITRs) into an 8-panel wizard form. The backend automatically parses inputs into structured JSON and populates AI suggested values.
High-level system topology showing interaction between the Next.js Frontend, FastAPI Backend, LlamaParse parsing, Elasticsearch / FAISS hybrid search, DeepSeek-V4 AI models, Document Vault, and Brevo Email service.
Deep dive into document ingestion, semantic chunking, metadata extraction, dual storage (SQLite3 + FAISS), hybrid BM25 + kNN vector search, Reciprocal Rank Fusion (RRF), and DeepSeek-V4 section classification.
The compilation lifecycle where classified section JSONs are validated, compiled by the Layout Engine, scored against SEBI compliance rules, rendered into final DRHP PDFs, and deployed to the ISO 27001 Document Vault for review and Brevo email sharing.
Full end-to-end master workflow diagram visualizing the entire ecosystem from document upload to final legal review.
🔍 Click to view Master Pipeline Architecture Diagram
IPO-Sahayak/
├── backend/ # Python FastAPI backend & ML microservices
│ ├── api/ # REST API endpoints & router setup
│ │ ├── __init__.py
│ │ ├── main.py # FastAPI application entry point & CORS configuration
│ │ └── routes.py # API routes for document upload, parsing & status polling
│ ├── config/ # Backend application configuration
│ │ ├── __init__.py
│ │ └── settings.py # App settings, environment variables & CORS settings
│ ├── ml/ # Machine Learning & NLP processing engines
│ │ ├── document_parser.py # Core PDF text extraction & document layout parser
│ │ ├── document_section_detector.py # Heuristic & ML section boundary detector
│ │ ├── section_classifier.py # NLP classifier mapping text to SEBI ICDR categories
│ │ ├── section_detector.py # Section segmentation & layout analyzer
│ │ └── main.py # Standalone ML pipeline test & runner script
│ ├── models/ # Data models & API request/response contracts
│ │ ├── __init__.py
│ │ └── schemas.py # Pydantic schemas for DRHP disclosures & financial metrics
│ ├── services/ # Business logic service layer
│ │ ├── __init__.py
│ │ ├── document_parser.py # High-level document parsing service wrapper
│ │ ├── email_service.py # Automated email notification & team invite dispatch
│ │ └── field_mapper.py # SEBI DRHP field mapping & schema converter
│ ├── requirements.txt # Python backend dependencies (FastAPI, PyPDF2, Pydantic, etc.)
│ └── .env # Backend environment variable configuration
│
├── jio_drhp_pipeline/ # Autonomous DRHP document synthesis & compilation pipeline
│ ├── main.py # Primary pipeline orchestration execution script
│ ├── config.py # AI LLM provider config (Fireworks AI, DeepSeek, OpenAI)
│ ├── check_setup.py # Environment verification & API setup test script
│ ├── requirements.txt # Python requirements for DRHP pipeline (requests, fitz, etc.)
│ ├── core/ # Core pipeline modular processing engines
│ │ ├── __init__.py
│ │ ├── classifier.py # Content categorization & prompt routing engine
│ │ ├── clients.py # AI LLM HTTP client wrappers
│ │ ├── compiler.py # LaTeX/PDF document compilation & rendering engine
│ │ ├── extractor.py # Financial & legal text extractor
│ │ ├── layout_engine.py # PDF visual layout & page composition manager
│ │ ├── retry.py # Resiliency, retry strategy & exponential backoff
│ │ ├── table_extractor.py # Structured table parser for balance sheets & P&L
│ │ ├── validator.py # SEBI ICDR compliance checker & rule validator
│ │ ├── vector_engine.py # RAG vector database & embedding manager
│ │ ├── versioning.py # Document version control & revision history
│ │ └── watermark.py # Draft watermarking & security overlay generator
│ └── sections_pipeline/ # SEBI Regulation specialized section generators
│ ├── 00_corporate_legal/ # Corporate Identity & Legal Disclosures Section
│ │ ├── prompts.py # AI prompts for legal background & promoter disclosures
│ │ ├── runner.py # Corporate legal section pipeline runner
│ │ └── schema.py # Data schemas for legal disclosures
│ ├── 04_financials/ # Financial Disclosures & Restated Statements
│ │ ├── prompts.py # AI prompts for financial restatement & ratios
│ │ ├── runner.py # Financial section runner & compiler
│ │ └── schema.py # Schemas for financial ratios & balance sheets
│ ├── 06_legal_regulatory/ # Litigation & Statutory Approvals
│ │ ├── prompts.py # Prompts for tax litigation & government permissions
│ │ ├── runner.py # Legal & regulatory pipeline runner
│ │ └── schema.py # Schema for litigation data
│ └── 07_capital_structure/ # Shareholding Pattern & Capital Structure
│ ├── prompts.py # AI prompts for pre/post-IPO capital structure
│ ├── runner.py # Capital structure calculator & table builder
│ └── schema.py # Schema for shareholding patterns & dilution
│
├── src/ # Next.js 15 App Router Frontend Application
│ ├── app/ # Next.js pages, layouts, and serverless API endpoints
│ │ ├── api/ # Next.js API Routes
│ │ │ ├── auth/ # User authentication & session endpoints
│ │ │ ├── companies/ # Company profile & MCA21 data endpoints
│ │ │ ├── send-document/ # Email & PDF dispatch handler
│ │ │ ├── translate/ # DeepSeek AI multi-language translation endpoint
│ │ │ └── uploads/ # Document upload API handler
│ │ ├── globals.css # Design tokens, custom glassmorphism & Tailwind styles
│ │ ├── layout.tsx # Root app layout with multi-language & toast providers
│ │ ├── page.tsx # Main Interactive Workspace Dashboard page
│ │ └── pdf-styles.css # CSS styles for live PDF document preview rendering
│ ├── components/ # Modular React UI Components
│ │ ├── Compliance/ # SEBI ICDR compliance tracker UI
│ │ │ └── ComplianceTracker.tsx # Interactive compliance checklist & rules engine
│ │ ├── Dashboard/ # Interactive analytics, team access & stats widgets
│ │ │ ├── CategoryChart.tsx # Visual charts for compliance categories
│ │ │ ├── Overview.tsx # Overview dashboard widget
│ │ │ ├── QuickStats.tsx # Quick metrics & stats cards
│ │ │ └── TeamAccess.tsx # RBAC team invitation & collaborator manager
│ │ ├── DocumentVault/ # Dual-panel PDF split viewer & document management
│ │ │ ├── CommentDrawer.tsx # Threaded commentary & review notes drawer
│ │ │ ├── SplitDocumentViewer.tsx # Side-by-side PDF preview & source document viewer
│ │ │ └── VaultManager.tsx # ISO 27001 document vault file manager
│ │ ├── GeneratedDocuments/ # Compiled DRHP document list & download manager
│ │ │ └── GeneratedDocuments.tsx # Generated PDF list, preview, & export actions
│ │ ├── Layout/ # Navigation header & dynamic sidebar layout
│ │ │ ├── Header.tsx # App header with language selector & profile controls
│ │ │ └── Sidebar.tsx # Collapsible navigation sidebar
│ │ ├── Onboarding/ # MCA21 CIN lookup & initial company setup wizard
│ │ │ └── Onboarding.tsx # Multi-step MCA21 CIN auto-fill onboarding flow
│ │ ├── UI/ # Reusable micro-interactive UI components
│ │ │ ├── CursorTrail.tsx # Custom cursor animation effect
│ │ │ ├── DocumentCard.tsx # Document card container
│ │ │ ├── EvidenceUpload.tsx # Document evidence upload dropzone
│ │ │ ├── FileUploadBox.tsx # Dynamic drag-and-drop file uploader
│ │ │ ├── HighlighterField.tsx # Source PDF text highlighter
│ │ │ ├── InkCheckbox.tsx # Styled custom checkbox component
│ │ │ ├── InkRadio.tsx # Styled custom radio button component
│ │ │ ├── MagneticButton.tsx # Interactive magnetic hover action button
│ │ │ ├── PaperStamp.tsx # Digital SEBI approval stamp animation
│ │ │ ├── ProgressIndicator.tsx # Custom progress ring & bar
│ │ │ ├── SignatureUploadRow.tsx # CA & Promoter digital signature uploader
│ │ │ ├── Toast.tsx # Custom notification toast alert
│ │ │ └── TypewriterText.tsx # Dynamic typing animation component
│ │ └── Wizard/ # 8-Panel DRHP step-by-step questionnaire system
│ │ ├── DocumentUploadZone.tsx # Financial & legal upload dropzone
│ │ ├── ExtractedDataPreview.tsx # OCR data extraction preview widget
│ │ ├── PanelOne_CorporateIdentity.tsx # Corporate details & promoter PAN info
│ │ ├── PanelTwo_BusinessOverview.tsx # Business model & Industry overview
│ │ ├── PanelThree_FinancialDossier.tsx # 3-Year financial metrics & restated financials
│ │ ├── PanelFour_MarketChannels.tsx # Market position, sales, & channels
│ │ ├── PanelFive_UseOfFundsLedger.tsx # Net IPO proceeds allocation & cap tables
│ │ ├── PanelSix_RiskAssessment.tsx # Internal & external risk factor builder
│ │ ├── PanelSeven_LegalDisclosures.tsx # Litigation, tax, & statutory permissions
│ │ ├── PanelEight_FinalReview.tsx # Final SEBI checklist & DRHP generation trigger
│ │ ├── StepFiveForm.tsx # Dynamic sub-form step renderer
│ │ ├── StepWizard.tsx # Wizard stepper layout manager & state wrapper
│ │ └── wizardTypes.ts # TypeScript definitions for wizard state & forms
│ ├── context/ # React Context State Providers
│ │ ├── GenerateDocumentsContext.tsx # Generation job status & progress state
│ │ ├── GeneratedDocumentsContext.tsx # Repository state for generated DRHP files
│ │ └── LanguageContext.tsx # Multi-language translation & UI dictionary state
│ ├── data/ # App state mocks & client interfaces
│ │ ├── generatedDocuments.ts # Sample generated DRHP document data
│ │ └── mockData.ts # Initial mock data for dashboard & compliance
│ ├── lib/ # Shared client-side helper utilities
│ │ ├── document.ts # Document processing helper functions
│ │ └── documents.ts # Client document management utilities
│ └── types/ # TypeScript type definitions
│ └── team-access.ts # Access control & RBAC role type definitions
│
├── data/ # Training data, parsed schemas, & document stores
│ ├── company1_prospectus_classified.json # SEBI classified DRHP sample dataset
│ ├── company1_prospectus_fixed_schema.json # Standardized JSON DRHP schema
│ ├── images/ # Extracted document graphics & charts
│ ├── markdown/ # Pre-parsed DRHP markdown documents
│ ├── pdfs/ # Source financial statement PDFs
│ └── spreadsheets/ # Sample financial statement Excel models
│
├── docs/ # Sample SME DRHPs & SEBI compliance reference documents
├── public/ # Static web assets & official logos
├── .env.local # Frontend environment variable configuration
├── eslint.config.mjs # ESLint configuration
├── next.config.ts # Next.js framework configuration
├── package.json # Node.js dependencies & script commands
├── postcss.config.mjs # PostCSS configuration for Tailwind CSS
├── requirements.txt # Root Python dependencies list
├── tsconfig.json # TypeScript compiler options
└── README.md # Comprehensive Project Documentation
Ensure your environment meets the following requirements:
- Node.js: v18.x or higher
- Python: 3.10 or higher
- Redis: v6.x or higher
- PostgreSQL: v14 or higher
- Elasticsearch: v8.x or higher
Configure the environment files for both backend and frontend applications.
Create a .env file in the backend/ directory:
# Database & Caching
DATABASE_URL=postgresql://user:password@localhost:5432/drhp_db
REDIS_URL=redis://localhost:6379/0
# Hybrid Vector Search (Elasticsearch)
ELASTICSEARCH_URL=http://localhost:9200
# AI & LLM Engine Keys
DEEPSEEK_API_KEY=your_deepseek_api_key
LLAMAPARSE_API_KEY=your_llamaparse_api_key
# Transactional Email & Notifications (Brevo API Integration)
BREVO_API_KEY=your_brevo_api_key
SENDER_EMAIL=notifications@iposahayak.com
SENDER_NAME="IPO Sahayak Platform"Create a .env.local file in the project root directory:
NEXT_PUBLIC_API_BASE_URL=http://localhost:8000
NEXT_PUBLIC_APP_URL=http://localhost:3000git clone https://github.com/Intelli2Byte/IPO-Sahayak-.git
cd IPO-Sahayak-# Install frontend dependencies
npm install
# Start development server
npm run devThe frontend application will be running at http://localhost:3000.
cd backend
# Create virtual environment
python -m venv venv
# Activate virtual environment
# On Linux/macOS:
source venv/bin/activate
# On Windows:
venv\Scripts\activate
# Install backend dependencies
pip install -r requirements.txt
# Launch FastAPI application server
uvicorn api.main:app --reload --port 8000The backend API documentation will be available at http://localhost:8000/docs.
cd backend
celery -A services.tasks worker --loglevel=info# Build frontend bundle
npm run buildWe welcome contributions from developers, AI engineers, legal compliance experts, and open-source enthusiasts! Follow these simple step-by-step guidelines to get started:
Fork the repository on GitHub and clone your fork locally:
git clone https://github.com/YOUR_USERNAME/IPO-Sahayak-.git
cd IPO-Sahayak-Create a descriptive branch for the feature or bugfix you are working on:
git checkout -b feature/amazing-new-featureInstall frontend dependencies and set up the Python backend virtual environment:
# Setup Frontend
npm install
# Setup Backend Virtual Environment
cd backend
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
pip install -r requirements.txtMake sure your changes adhere to project code style standards:
git add .
git commit -m "feat(ai): integrate DeepSeek-V4 section validation rules"Push your feature branch to your GitHub repository and submit a Pull Request:
git push origin feature/amazing-new-feature- Ensure all TypeScript files pass linting (
npm run lint). - Ensure Python scripts conform to PEP8 standards.
- Include concise commit messages explaining the rationale behind your changes.
- Tag related issue numbers in your Pull Request description.
Distributed under the MIT License. See LICENSE for details. Built to democratize public capital market access for Indian SMEs.