Raniux

Raniux - the only market research tool you'll ever need
Find all answers via a chat interface
Tap into highest level of subject matter expertise
Find competitors with precision
Experience a completely personalized experience

Inspiration

As a venture capital investor, I've spent years dreaming of an intelligent partner—a fully-fledged technology and investment AI that could act as a true extension of my own analytical process. I wanted a tool that didn't just find data, but synthesized it into strategic insight; something that could test a thesis, map a market, and diligence a company with the speed and depth I always wished for. Raniux is the realization of that dream.

What it does

Raniux is an enterprise-grade market intelligence platform designed for investors, strategists, and founders. It leverages a team of collaborative AI agents to automate deep market research, perform company due diligence, and generate sophisticated reports. Users can chat with the platform in real-time, get instant answers from a production-ready RAG system, and receive unparalleled strategic insights that were previously impossible to obtain without a dedicated team of human analysts.

How we built it

We architected Raniux using a modern, scalable stack designed to use the best tool for each job.

UX & Frontend: The entire user experience was built rapidly using Bolt.

Backend-as-a-Service: We used Supabase to handle the database, authentication, and 33 specialized Edge Functions that form our business logic microservices.

AI Agent Processing: Because our sophisticated AI agents require Python and involve long-running processes, we built them as containerized applications on GCP Cloud Run. This created a robust, hybrid system where lightweight edge functions can trigger and poll these powerful agentic workflows.

Complex Chat Logic: To handle the intricate, stateful logic of our streaming chat interface, we used Cursor as an AI-assisted development environment. This allowed us to build a "fat edge function" with over 2,000 lines of highly optimized code, responsible for everything from context management to real-time intent classification.

Challenges we ran into

Our primary challenge was architectural: integrating fast, serverless edge functions with the long-running Python processes required for our AI agents. A pure Bolt and Supabase stack can't natively support these intensive tasks.

We overcame this by designing a hybrid cloud system where Supabase orchestrates the jobs, passing them to GCP Cloud Run for heavy lifting and then polling for the results. This required careful engineering to ensure seamless communication, robust error handling, and a smooth user experience, effectively bridging the gap between two powerful but distinct cloud environments.

Accomplishments that we're proud of

It feels like I was finally able to build the app I was trying to build with a 4-person engineering team for 18 months with little progress! We are incredibly proud of architecting and delivering a complete, enterprise-grade platform in such a short time.

Key accomplishments include:

A full, production-ready RAG system with gte-small embeddings powering our semantic company search.

A flexible, real-time streaming chat interface with sophisticated routing for several user intents, from simple Q&A to triggering complex reports.

The multi-agent report generation system, capable of producing professionally styled, data-rich PDF reports on demand.

Achieving true platform-level integration, where a user can seamlessly move from a chat conversation to a deep-dive report, powered by a combination of RAG, agentic workflows, and hybrid search.

What we learned

This project was a masterclass in modern AI application development. We learned how to bring a complex ecosystem of technologies together:

Orchestrating AI agents in a hybrid cloud environment.

Implementing a hybrid search system that combines semantic and keyword strengths.

Building a production-ready RAG pipeline from the ground up.

Designing and routing user intent from a single chat interface.

On a practical level, we honed our skills in navigating a complex tech stack, implementing rigorous debugging and testing routines, and syncing our work effectively through GitHub, preparing us to run this as a production system in a matter of weeks.

What's next for Raniux

Our path forward is clear and focused:

Begin extensive platform testing with synthetic users to identify and resolve edge cases.
Move to closed beta testing with a curated group of real target users to gather critical feedback.
Approach our 1,500+ investor waitlist, providing each user with a personalized onboarding experience designed to demonstrate immediate value and convert them into our first wave of paying customers.