Platform Documentation
Understand the advanced Retrieval-Augmented Generation models, federated crawlers, and physics maps powering Forfwd.
# 1. Spin up the sandboxed local SearXNG container
docker compose up -d
# 2. Install all core packages & dependencies
npm install
# 3. Fire up the high-performance Next.js dev server
npm run dev1. Retrieval-Augmented Generation (RAG)
Forfwd uses a highly optimized, multi-agent RAG pipeline rather than relying on static pre-trained weights. When you enter a career or academic question, Forfwd dynamically synthesizes search queries, scrapes live job markets and university criteria via federated SearXNG containers, cleans raw HTML contexts, and feeds validated contexts into Google Gemini 1.5 Flash inside deterministic Zod schemas.
2. Real-Time Web Synthesis (SearXNG)
SearXNG serves as Forfwd's distributed retrieval search gateway. Running in a sandboxed docker container, it queries up to 10 endpoints concurrently. This parallel execution minimizes latency and completely avoids search rate-limits, providing live, halluncination-free data sources (degree costs, market demand, hiring trends) within milliseconds.
3. Interactive Physics Orbit Map
Our signature 'VisualExplorer' utilizes custom HTML5 canvas node layouts to represent alternative career trajectories and educational requirements. Each path is mapped as an orbit around a student's core specialty, featuring realistic hover-reactive animations, friction physics, and dynamic connection lines.
4. ATS Resume Optimization
Our ATS Engine parses resume contents and checks them against live job postings to identify critical skill gaps. Our AI-driven resume rewriter suggests seamless contextual integrations of these missing entities without disrupting the original tone, providing an interactive split-screen diff viewer for simple code copy-paste.
Send us a direct message
Have questions about compliance, technical RAG criteria, or course tracking? Fill out the form below to save your inquiry directly into our secure Neon database.